Skip to content

Chip (Institutional)

In Taiwan stock chip data, we have 24 datasets as follows:

Individual Stock Margin Purchase / Short Sale TaiwanStockMarginPurchaseShortSale

  • Data range: 2001-01-01 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_margin_purchase_short_sale(
    stock_id="2330",
    start_date='2020-04-02',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockMarginPurchaseShortSale",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockMarginPurchaseShortSale",
        data_id= "2330",
        start_date= "2020-01-02",
        end_date= "2020-04-12",
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_margin_purchase_short_sale(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

date stock_id MarginPurchaseBuy MarginPurchaseCashRepayment MarginPurchaseLimit MarginPurchaseSell MarginPurchaseTodayBalance MarginPurchaseYesterdayBalance Note OffsetLoanAndShort ShortSaleBuy ShortSaleCashRepayment ShortSaleLimit ShortSaleSell ShortSaleTodayBalance ShortSaleYesterdayBalance
0 2020-04-06 2330 1914 8 6482595 1269 26285 25648 X 0 0 24 6482595 0 0 24
1 2020-04-07 2330 1049 13 6482595 2655 24666 26285 X 0 0 0 6482595 0 0 0
2 2020-04-08 2330 1192 3 6482595 1569 24286 24666 0 0 0 6482595 0 0 0
3 2020-04-09 2330 499 28 6482595 1362 23395 24286 209 0 0 6482595 398 398 0
4 2020-04-10 2330 1227 24 6482595 794 23804 23395 53 156 0 6482595 156 398 398 
{
    date: str, # date
    stock_id: str, # stock symbol
    MarginPurchaseBuy: int64, # margin purchase buy
    MarginPurchaseCashRepayment: int64, # margin purchase cash repayment
    MarginPurchaseLimit: int64, # margin purchase limit
    MarginPurchaseSell: int64, # margin purchase sell
    MarginPurchaseTodayBalance: int64, # margin purchase today balance
    MarginPurchaseYesterdayBalance: int64, # margin purchase yesterday balance
    Note: str, # note
    OffsetLoanAndShort: int64, # offset loan and short
    ShortSaleBuy: int64, # short sale buy
    ShortSaleCashRepayment: int64, # short sale cash repayment
    ShortSaleLimit: int64, # short sale limit
    ShortSaleSell: int64, # short sale sell
    ShortSaleTodayBalance: int64, # short sale today balance
    ShortSaleYesterdayBalance: int64 # short sale yesterday balance
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_margin_purchase_short_sale(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockMarginPurchaseShortSale",
    "start_date": "2020-04-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockMarginPurchaseShortSale",
        start_date= "2020-01-02"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id MarginPurchaseBuy MarginPurchaseCashRepayment MarginPurchaseLimit MarginPurchaseSell MarginPurchaseTodayBalance MarginPurchaseYesterdayBalance Note OffsetLoanAndShort ShortSaleBuy ShortSaleCashRepayment ShortSaleLimit ShortSaleSell ShortSaleTodayBalance ShortSaleYesterdayBalance
0 2020-04-01 0050 193 15 263750 163 3189 3174 0 65 1 263750 13 2283 2336
1 2020-04-01 0051 0 0 2375 0 5 5 0 0 0 2375 0 0 0
2 2020-04-01 0052 0 0 7500 0 128 128 0 0 0 7500 0 0 0
3 2020-04-01 0053 0 0 1622 0 1 1 0 0 0 1622 0 0 0
4 2020-04-01 0054 0 0 2531 0 0 0 X 0 0 0 2531 0 0 0 
{
    date: str, # date
    stock_id: str, # stock symbol
    MarginPurchaseBuy: int64, # margin purchase buy
    MarginPurchaseCashRepayment: int64, # margin purchase cash repayment
    MarginPurchaseLimit: int64, # margin purchase limit
    MarginPurchaseSell: int64, # margin purchase sell
    MarginPurchaseTodayBalance: int64, # margin purchase today balance
    MarginPurchaseYesterdayBalance: int64, # margin purchase yesterday balance
    Note: str, # note
    OffsetLoanAndShort: int64, # offset loan and short
    ShortSaleBuy: int64, # short sale buy
    ShortSaleCashRepayment: int64, # short sale cash repayment
    ShortSaleLimit: int64, # short sale limit
    ShortSaleSell: int64, # short sale sell
    ShortSaleTodayBalance: int64, # short sale today balance
    ShortSaleYesterdayBalance: int64 # short sale yesterday balance
}

Taiwan Total Market Margin Purchase / Short Sale TaiwanStockTotalMarginPurchaseShortSale

  • Data range: 2001-01-01 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_margin_purchase_short_sale_total(
    start_date='2020-04-01',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockTotalMarginPurchaseShortSale",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockTotalMarginPurchaseShortSale",
        start_date= "2020-01-02",
        end_date= "2020-04-12"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

TodayBalance YesBalance buy date name Return sell
0 5463820 5471770 236127 2020-04-01 MarginPurchase 10986 233091
1 91965082000 91898116000 4046643000 2020-04-01 MarginPurchaseMoney 196619000 3783058000
2 541704 556742 57266 2020-04-01 ShortSale 6151 48379
3 535401 541704 50779 2020-04-06 ShortSale 3700 48176
4 93198509000 91965082000 6440842000 2020-04-06 MarginPurchaseMoney 71638000 5135777000 
{
    TodayBalance: int64, # today balance
    YesBalance: int64, # yesterday balance
    buy: int64, # buy
    date: str, # date
    name: str, # category
    Return: int64, # cash/securities repayment
    sell: int64 # sell
}

Institutional Investors Buy/Sell TaiwanStockInstitutionalInvestorsBuySell

  • Data range: 2005-01-01 ~ now
  • Coverage: listed (TWSE), OTC (TPEx) and emerging market companies, distinguished by stock_id (use TaiwanStockInfo to look up market type)
  • Data update time Monday to Friday 20:00, the actual update time is based on the API data.
Values may be updated slightly after the initial release

This data is the original-trade statistics of each institutional investor for the day. A small number of trades (e.g. after-hours fixed-price or block/matched trades) may be tallied into the day's statistics only after the initial release, so a given day's buy/sell totals can be topped up afterwards; when both sides of such a trade belong to the same investor category, buy and sell increase by the same amount and the net (buy minus sell) is unaffected. This is the completion of the day's original trades — not a data error, and not an error-account correction. There is no fixed "finalized" time; treat the value currently returned by the API as the latest. If you need finalized figures, retrieve them one or two business days after the trading day.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_institutional_investors(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockInstitutionalInvestorsBuySell",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockInstitutionalInvestorsBuySell",
        data_id= "2330",
        start_date= "2020-04-01",
        end_date= "2020-04-12"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_institutional_investors(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

date stock_id buy name sell
0 2020-04-01 2330 31304729 Foreign_Investor 29057663
1 2020-04-01 2330 0 Foreign_Dealer_Self 0
2 2020-04-01 2330 900000 Investment_Trust 239000
3 2020-04-01 2330 79000 Dealer_self 807000
4 2020-04-01 2330 189000 Dealer_Hedging 493500 
{
    date: str, # date
    stock_id: str, # stock symbol
    buy: int64, # buy
    name: str, # category
    sell: int64 # sell
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_institutional_investors(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockInstitutionalInvestorsBuySell",
    "start_date": "2020-04-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data)

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockInstitutionalInvestorsBuySell",
        start_date= "2020-01-02"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id buy name sell
0 2020-04-01 0050 458249 Foreign_Investor 4492000
1 2020-04-01 0050 0 Foreign_Dealer_Self 0
2 2020-04-01 0050 54000 Investment_Trust 0
3 2020-04-01 0050 0 Dealer_self 0
4 2020-04-01 0050 2050000 Dealer_Hedging 905000 
{
    date: str, # date
    stock_id: str, # stock symbol
    buy: int64, # buy
    name: str, # category
    sell: int64 # sell
}

Institutional Investors Buy/Sell (Wide) TaiwanStockInstitutionalInvestorsBuySellWide

  • Same source data as TaiwanStockInstitutionalInvestorsBuySell, but in wide format: one row per trading day, with each institutional investor's buy/sell flattened into its own column — no manual pivot needed.
  • Data range: 2005-01-01 ~ now
  • Coverage: listed (TWSE), OTC (TPEx) and emerging market companies, distinguished by stock_id (use TaiwanStockInfo to look up market type)
  • Data update time Monday to Friday 20:00, the actual update time is based on the API data.
Values may be updated slightly after the initial release

This data is the original-trade statistics of each institutional investor for the day. A small number of trades (e.g. after-hours fixed-price or block/matched trades) may be tallied into the day's statistics only after the initial release, so a given day's buy/sell totals can be topped up afterwards; when both sides of such a trade belong to the same investor category, buy and sell increase by the same amount and the net (buy minus sell) is unaffected. This is the completion of the day's original trades — not a data error, and not an error-account correction. There is no fixed "finalized" time; treat the value currently returned by the API as the latest. If you need finalized figures, retrieve them one or two business days after the trading day.

New vs old era columns (important)

The classification of the three institutional investors has changed over the years. This wide table includes all historical classifications; a classification is always 0 in eras where it did not yet exist:

Column Available era Other eras
Foreign_Investor (foreign, excl. foreign dealer self) all —
Investment_Trust all —
Dealer (dealer, combined) old era (before ~2014-12-01) and Emerging 0 afterwards
Dealer_self (dealer, proprietary) new era (from 2014-12-01) 0 before
Dealer_Hedging (dealer, hedging) new era (from 2014-12-01) 0 before
Foreign_Dealer_Self (foreign dealer self) new era (from 2018-01-15) 0 before
  • From 2014-12-01, the combined Dealer was split into Dealer_self (proprietary) and Dealer_Hedging (hedging).
  • From 2018-01-15, Foreign_Dealer_Self was split out from foreign investors.
  • For a continuous "dealer total" across eras, sum Dealer + Dealer_self + Dealer_Hedging (only one group is non-zero in any era, so there is no double counting).

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_institutional_investors_wide(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # login to get the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockInstitutionalInvestorsBuySellWide",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # login to get the token
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockInstitutionalInvestorsBuySellWide",
        data_id= "2330",
        start_date= "2020-04-01",
        end_date= "2020-04-12"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id Foreign_Investor_buy Foreign_Investor_sell Foreign_Dealer_Self_buy Foreign_Dealer_Self_sell Investment_Trust_buy Investment_Trust_sell Dealer_buy Dealer_sell Dealer_self_buy Dealer_self_sell Dealer_Hedging_buy Dealer_Hedging_sell
0 2020-04-01 2330 31304729 29057663 0 0 900000 239000 0 0 79000 807000 189000 493500 
{
    date: str, # date
    stock_id: str, # stock id
    Foreign_Investor_buy: int64, # foreign investor buy
    Foreign_Investor_sell: int64, # foreign investor sell
    Foreign_Dealer_Self_buy: int64, # foreign dealer self buy
    Foreign_Dealer_Self_sell: int64, # foreign dealer self sell
    Investment_Trust_buy: int64, # investment trust buy
    Investment_Trust_sell: int64, # investment trust sell
    Dealer_buy: int64, # dealer buy (combined, old era)
    Dealer_sell: int64, # dealer sell (combined, old era)
    Dealer_self_buy: int64, # dealer proprietary buy
    Dealer_self_sell: int64, # dealer proprietary sell
    Dealer_Hedging_buy: int64, # dealer hedging buy
    Dealer_Hedging_sell: int64 # dealer hedging sell
}

Taiwan Total Market Institutional Investors Buy/Sell TaiwanStockTotalInstitutionalInvestors

  • Data range: 2004-04-01 ~ now
  • Data update time Monday to Friday 15:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_institutional_investors_total(
    start_date='2020-04-01',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockTotalInstitutionalInvestors",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockTotalInstitutionalInvestors",
        start_date= "2020-01-02",
        end_date='2020-04-12'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

buy date name sell
0 123150 2020-04-01 Foreign_Dealer_Self 266220
1 3681729831 2020-04-01 Dealer_Hedging 5539788946
2 33759089839 2020-04-01 Foreign_Investor 38466572585
3 3039112340 2020-04-01 Investment_Trust 853138940
4 789316840 2020-04-01 Dealer_self 912143500 
{
    buy: int64, # buy
    date: str, # date
    name: str, # category
    sell: int64 # sell
}

Foreign Investor Shareholding TaiwanStockShareholding

  • Data range: 2004-02-01 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_shareholding(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12'
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockShareholding",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockShareholding",
        data_id= "2330",
        start_date= "2020-01-02",
        end_date="2020-04-12"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_shareholding(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

date stock_id stock_name InternationalCode ForeignInvestmentRemainingShares ForeignInvestmentShares ForeignInvestmentRemainRatio ForeignInvestmentSharesRatio ForeignInvestmentUpperLimitRatio ChineseInvestmentUpperLimitRatio NumberOfSharesIssued RecentlyDeclareDate note
0 2020-04-01 2330 台積電 TW0002330008 6309042842 19621337616 24.33 75.66 100 100 25930380458 2019-05-27
1 2020-04-06 2330 台積電 TW0002330008 6304552683 19625827775 24.31 75.68 100 100 25930380458 2019-05-27
2 2020-04-07 2330 台積電 TW0002330008 6283562246 19646818212 24.23 75.76 100 100 25930380458 2019-05-27
3 2020-04-08 2330 台積電 TW0002330008 6273338931 19657041527 24.19 75.8 100 100 25930380458 2019-05-27
4 2020-04-09 2330 台積電 TW0002330008 6267988722 19662391736 24.17 75.82 100 100 25930380458 2019-05-27 
{
    date: str, # date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    InternationalCode: str, # international stock code
    ForeignInvestmentRemainingShares: int64, # remaining shares available for foreign investment
    ForeignInvestmentShares: int64, # shares held by foreign investors
    ForeignInvestmentRemainRatio: float64, # ratio of shares available for foreign investment
    ForeignInvestmentSharesRatio: float64, # foreign shareholding ratio
    ForeignInvestmentUpperLimitRatio: float64, # foreign investment upper limit
    ChineseInvestmentUpperLimitRatio: float64, # mainland Chinese investment upper limit
    NumberOfSharesIssued: int64, # number of shares issued
    RecentlyDeclareDate: str, # most recent declaration date
    note: str # note
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_shareholding(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockShareholding",
    "start_date": "2020-04-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockShareholding",
        start_date= "2020-01-02"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id stock_name InternationalCode ForeignInvestmentRemainingShares ForeignInvestmentShares ForeignInvestmentRemainRatio ForeignInvestmentSharesRatio ForeignInvestmentUpperLimitRatio ChineseInvestmentUpperLimitRatio NumberOfSharesIssued RecentlyDeclareDate note
0 2020-04-01 0050 元大台灣50 TW0000050004 960256795 94743205 91.01 8.98 100 100 1055000000 2019-07-18
1 2020-04-01 0051 元大中型100 TW0000051002 9471000 29000 99.69 0.3 100 100 9500000 2019-07-18
2 2020-04-01 0052 富邦科技 TW0000052000 29957000 43000 99.85 0.14 100 100 30000000 2019-07-18
3 2020-04-01 0053 元大電子 TW0000053008 6466950 21050 99.67 0.32 100 100 6488000 2019-07-18
4 2020-04-01 0054 元大台商50 TW0000054006 9955000 169000 98.33 1.66 100 100 10124000 2019-07-18 
{
    date: str, # date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    InternationalCode: str, # international stock code
    ForeignInvestmentRemainingShares: int64, # remaining shares available for foreign investment
    ForeignInvestmentShares: int64, # shares held by foreign investors
    ForeignInvestmentRemainRatio: float64, # ratio of shares available for foreign investment
    ForeignInvestmentSharesRatio: float64, # foreign shareholding ratio
    ForeignInvestmentUpperLimitRatio: float64, # foreign investment upper limit
    ChineseInvestmentUpperLimitRatio: float64, # mainland Chinese investment upper limit
    NumberOfSharesIssued: int64, # number of shares issued
    RecentlyDeclareDate: str, # most recent declaration date
    note: str # note
}

Shareholders Holding Shares Distribution TaiwanStockHoldingSharesPer (only available for backer, sponsor members)

  • Data range: 2010-01-29 ~ now

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_holding_shares_per(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12'
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockHoldingSharesPer",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockHoldingSharesPer",
        data_id= "2330",
        start_date= "2020-01-02",
        end_date='2020-04-12'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_holding_shares_per(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

date stock_id HoldingSharesLevel people percent unit
0 2020-04-01 2330 1-999 165122 0.12 33289900
1 2020-04-01 2330 1000-5000 227692 1.69 440404454
2 2020-04-01 2330 10001-15000 10408 0.49 128127693
3 2020-04-01 2330 100001-200000 1628 0.86 225202876
4 2020-04-01 2330 15001-20000 5068 0.34 89929303 
{
    date: str, # date
    stock_id: str, # stock symbol
    HoldingSharesLevel: str, # shareholding level
    people: int64, # number of people
    percent: float64, # ratio
    unit: int64 # number of shares
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_holding_shares_per(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockHoldingSharesPer",
    "start_date": "2020-04-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data)

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockHoldingSharesPer",
        start_date= "2020-04-01"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id HoldingSharesLevel people percent unit
0 2020-04-01 0050 1-999 44173 1.02 10834763
1 2020-04-01 0050 1000-5000 96465 17.7 186791648
2 2020-04-01 0050 5001-10000 10364 7.57 79902735
3 2020-04-01 0050 10001-15000 2819 3.41 36075583
4 2020-04-01 0050 15001-20000 1557 2.69 28426726 
{
    date: str, # date
    stock_id: str, # stock symbol
    HoldingSharesLevel: str, # shareholding level
    people: int64, # number of people
    percent: float64, # ratio
    unit: int64 # number of shares
}

Securities Lending Transaction Details TaiwanStockSecuritiesLending

  • Data range: 2001-05-01 ~ now
  • Data update time Monday to Friday 15:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_securities_lending(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12'
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockSecuritiesLending",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
url = url,
    query = list(
        dataset="TaiwanStockSecuritiesLending",
        data_id="2330",
        start_date= "2020-01-02",
        end_date='2020-04-12'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_securities_lending(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

date stock_id transaction_type volume fee_rate close original_return_date original_lending_period
0 2020-04-01 2330 議借 1330 1.36 271.5 2020-09-30 182
1 2020-04-01 2330 議借 800 0.41 271.5 2020-09-30 182
2 2020-04-01 2330 議借 850 0.41 271.5 2020-09-30 182
3 2020-04-01 2330 議借 500 0.5 271.5 2020-09-30 182
4 2020-04-01 2330 議借 160 0.36 271.5 2020-09-30 182 
{
    date: str, # date
    stock_id: str, # stock symbol
    transaction_type: str, # transaction type
    volume: int64, # transaction volume
    fee_rate: float64, # transaction fee rate
    close: float64, # closing price
    original_return_date: str, # agreed return date
    original_lending_period: int64 # agreed lending period (days)
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_securities_lending(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockSecuritiesLending",
    "start_date": "2020-04-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data)

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockSecuritiesLending",
        start_date= "2020-01-02"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id transaction_type volume fee_rate close original_return_date original_lending_period
0 2020-04-01 1101 議借 760 0.25 39 2020-09-30 182
1 2020-04-01 1101 議借 397 0.25 39 2020-09-30 182
2 2020-04-01 1101 競價 436 0.7 39 2020-09-30 182
3 2020-04-01 1102 議借 150 0.25 38.6 2020-09-30 182
4 2020-04-01 1102 議借 770 1.05 38.6 2020-09-30 182 
{
    date: str, # date
    stock_id: str, # stock symbol
    transaction_type: str, # transaction type
    volume: int64, # transaction volume
    fee_rate: float64, # transaction fee rate
    close: float64, # closing price
    original_return_date: str, # agreed return date
    original_lending_period: int64 # agreed lending period (days)
}

Margin Short Sale Suspension (Short Sale Covering Date) TaiwanStockMarginShortSaleSuspension

  • Data range: 2015-01-01 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockMarginShortSaleSuspension",
    "data_id": "0050",
    "start_date": "2015-01-01",
    "end_date": "2015-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockMarginShortSaleSuspension",
        data_id="0050",
        start_date= "2015-01-01",
        end_date= "2015-04-12"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_stock_margin_short_sale_suspension(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

stock_id date end_date reason
0 0050 2015-10-20 2015-10-23 分配收益
1 0050 2016-07-22 2016-07-27 分配收益
2 0050 2017-02-02 2017-02-07 分配收益
3 0050 2017-07-25 2017-07-28 分配收益
4 0050 2018-01-23 2018-01-26 分配收益 
{
    stock_id: str, # stock symbol
    date: str, # start date
    end_date: str, # end date
    reason: str # suspension reason
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockMarginShortSaleSuspension",
    "start_date": "2015-10-20",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data)

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockMarginShortSaleSuspension",
        start_date= "2015-10-20"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

stock_id date end_date reason
0 0050 2015-10-20 2015-10-23 分配收益
1 0056 2015-10-20 2015-10-23 分配收益 
{
    stock_id: str, # stock symbol
    date: str, # start date
    end_date: str, # end date
    reason: str # suspension reason
}

Credit Limit Daily Short Sale Balances TaiwanDailyShortSaleBalances

  • Data range: 2005-07-01 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_daily_short_sale_balances(
    stock_id="2330",
    start_date='2020-04-01',
    end_date='2020-04-12',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanDailyShortSaleBalances",
    "data_id": "2330",
    "start_date": "2020-04-01",
    "end_date": "2020-04-12",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
url = url,
    query = list(
        dataset="TaiwanDailyShortSaleBalances",
        data_id="2330",
        start_date= "2020-01-02",
        end_date='2020-04-12'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

start = datetime.datetime.now()
df = api.taiwan_daily_short_sale_balances(
    stock_id_list=['2330', '2317', '2454', '3008'],
    start_date='2024-01-01',
    end_date='2024-12-31',
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

stock_id MarginShortSalesPreviousDayBalance MarginShortSalesShortSales MarginShortSalesShortCovering MarginShortSalesStockRedemption MarginShortSalesCurrentDayBalance MarginShortSalesQuota SBLShortSalesPreviousDayBalance SBLShortSalesShortSales SBLShortSalesReturns SBLShortSalesAdjustments SBLShortSalesCurrentDayBalance SBLShortSalesQuota SBLShortSalesShortCovering date
0 2330 1975000 0 1573000 378000 24000 -2107339478 47947858 487000 0 0 48434858 7526895 0 2020-04-01
1 2330 24000 0 0 24000 0 -2107339478 48434858 44000 60000 0 48418858 7563083 0 2020-04-06
2 2330 0 0 0 0 0 -2107339478 48418858 62000 0 0 48480858 7635835 0 2020-04-07
3 2330 0 0 0 0 0 -2107339478 48480858 933000 7345000 0 42068858 7688249 0 2020-04-08
4 2330 0 398000 0 0 398000 -2107339478 42068858 46000 2000 0 42112858 7642682 0 2020-04-09 
{
    stock_id: str, # stock symbol
    MarginShortSalesPreviousDayBalance: int32, # previous day balance (margin short sales)
    MarginShortSalesShortSales: int32, # sell (margin short sales)
    MarginShortSalesShortCovering: int32, # buy (margin short sales)
    MarginShortSalesStockRedemption: int32, # stock redemption (margin short sales)
    MarginShortSalesCurrentDayBalance: int32, # current day balance (margin short sales)
    MarginShortSalesQuota: int32, # quota (margin short sales)
    SBLShortSalesPreviousDayBalance: int32, # previous day balance (SBL short sales)
    SBLShortSalesShortSales: int32, # sell (SBL short sales)
    SBLShortSalesReturns: int32, # returns (SBL short sales)
    SBLShortSalesAdjustments: int32, # current day adjustment (SBL short sales)
    SBLShortSalesCurrentDayBalance: int32, # current day balance (SBL short sales)
    SBLShortSalesQuota: int32, # quota (SBL short sales)
    SBLShortSalesShortCovering: int32, # inventory adjustment (SBL short sales)
    date: str # date
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_daily_short_sale_balances(
    start_date='2020-04-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanDailyShortSaleBalances",
    "start_date": "2021-05-20",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data)

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
url = url,
    query = list(
        dataset="TaiwanDailyShortSaleBalances",
        start_date= "2020-01-02"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

stock_id MarginShortSalesPreviousDayBalance MarginShortSalesShortSales MarginShortSalesShortCovering MarginShortSalesStockRedemption MarginShortSalesCurrentDayBalance MarginShortSalesQuota SBLShortSalesPreviousDayBalance SBLShortSalesShortSales SBLShortSalesReturns SBLShortSalesAdjustments SBLShortSalesCurrentDayBalance SBLShortSalesQuota SBLShortSalesShortCovering date
0 0050 2336000 13000 65000 1000 2283000 263750000 25527000 0 0 0 25527000 2397551 0 2020-04-01
1 0051 0 0 0 0 0 2375000 1000 0 0 0 1000 4053 0 2020-04-01
2 0052 0 0 0 0 0 7500000 34000 0 0 0 34000 17168 0 2020-04-01
3 0053 0 0 0 0 0 1622000 0 0 0 0 0 3158 0 2020-04-01
4 0054 0 0 0 0 0 2531000 0 0 0 0 0 1357 0 2020-04-01 
{
    stock_id: str, # stock symbol
    MarginShortSalesPreviousDayBalance: int32, # previous day balance (margin short sales)
    MarginShortSalesShortSales: int32, # sell (margin short sales)
    MarginShortSalesShortCovering: int32, # buy (margin short sales)
    MarginShortSalesStockRedemption: int32, # stock redemption (margin short sales)
    MarginShortSalesCurrentDayBalance: int32, # current day balance (margin short sales)
    MarginShortSalesQuota: int32, # quota (margin short sales)
    SBLShortSalesPreviousDayBalance: int32, # previous day balance (SBL short sales)
    SBLShortSalesShortSales: int32, # sell (SBL short sales)
    SBLShortSalesReturns: int32, # returns (SBL short sales)
    SBLShortSalesAdjustments: int32, # current day adjustment (SBL short sales)
    SBLShortSalesCurrentDayBalance: int32, # current day balance (SBL short sales)
    SBLShortSalesQuota: int32, # quota (SBL short sales)
    SBLShortSalesShortCovering: int32, # inventory adjustment (SBL short sales)
    date: str # date
}

Securities Trader Information TaiwanSecuritiesTraderInfo

  • Provides securities trader related information, used with the Taiwan Stock Trading Daily Report by Branch (TaiwanStockTradingDailyReport). Use the securities_trader_id to query all stock transactions of a particular securities trader.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_securities_trader_info()

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanSecuritiesTraderInfo",
}
resp = requests.get(url, headers=headers, params=parameter)
data = resp.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset = "TaiwanSecuritiesTraderInfo"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

securities_trader_id securities_trader date address phone
0 1020 合庫 2011-12-02 台北市大安區忠孝東路四段325號2樓(部分)、經紀部複委託科地址：台北市松山區長安東路二段225號5樓 02-27528000
1 1021 合庫- 台中 2011-12-02 台中市西區民權路91號6樓 04-22255141
2 1022 合庫-台南 2011-12-02 台南市北區成功路48號3樓 06-2260148
3 1023 合庫-高雄 2011-12-02 高雄市大勇路97號5樓 07-5319755
4 1024 合庫-嘉義 2011-12-02 嘉義市國華街279號2樓 05-2220016 
{
    securities_trader_id: str, # securities trader code
    securities_trader: str, # securities trader name
    date: str, # opening date
    address: str, # address
    phone: str # phone
}

Taiwan Stock Trading Daily Report by Branch (query by stock_id) TaiwanStockTradingDailyReport (only available for sponsor members)

  • Provides Taiwan stock trading by branch information for listed (TWSE), OTC, and emerging stocks!
  • Data range: 2021-06-30 ~ now
  • Due to the large data volume, only one day of data is provided per request.
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.
  • Some data is missing on the following dates: 2022-10-31~2022-11-03, 2023-01-11~2023-01-17.
  • Enabling Async can significantly reduce data fetch time. In Colab tests, downloading 2,175 stocks takes only 4 minutes 20 seconds.
A price of 0 on emerging-stock dealer (自營商) branches is expected, not missing data

Emerging stocks (where TaiwanStockInfo reports the stock_id as emerging) trade on a quote-driven negotiation market run by recommending securities firms, unlike the central-matching auction used for listed/OTC stocks. The recommending firms act as market makers, continuously quoting bid/ask and filling orders with their own capital and inventory — i.e. the dealer (自營商) branch, whose securities_trader_id ends in T. Because that daily market-making position trades both sides across many price levels, there is no single representative execution price, so price is reported as 0 for that branch row (buy/sell share counts are still correct).

  • This is a structural characteristic of the emerging market, not missing or erroneous data; ordinary (non-dealer) branches of the same emerging stock still carry normal execution prices.
  • Listed (TWSE) and OTC stocks are unaffected — every branch, including dealers, has an execution price.
  • Reference: TPEx — Emerging Stock Trading System.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_trading_daily_report(
    stock_id="2330",
    date='2022-06-16',
)
print(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

token = ""

data_loader = DataLoader()
data_loader.login_by_token(token)

date = "2025-12-08"
start = datetime.datetime.now()
df = data_loader.taiwan_stock_trading_daily_report(
    date=date,
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)
# 0:04:20

import requests
import pandas as pd

url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report'
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "data_id": "2330",
    "date": "2022-06-16",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
token = "" # Refer to login to obtain the API key
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report'
response = httr::GET(
    url = url,
    query = list(
        data_id="2330",
        start_date= "2022-06-16",
        token = token # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 合庫 508 4000 2000 1020 2330 2022-06-16
1 合庫 509 3480 0 1020 2330 2022-06-16
2 合庫 510 2310 50 1020 2330 2022-06-16
3 合庫 511 1169 0 1020 2330 2022-06-16
4 合庫 512 1300 10000 1020 2330 2022-06-16 
{
    securities_trader: str, # securities trader name
    price: float64, # transaction price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Taiwan Stock Trading Daily Report by Branch (query by securities_trader_id) TaiwanStockTradingDailyReport (only available for sponsor members)

  • Provides Taiwan stock trading by branch information for listed (TWSE), OTC, and emerging stocks!
  • Data range: 2021-06-30 ~ now
  • Due to the large data volume, only one day of data is provided per request.
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.
  • Some data is missing on the following dates: 2022-10-31~2022-11-03, 2023-01-11~2023-01-17.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_trading_daily_report(
    securities_trader_id="1102",
    date='2022-06-16',
)
print(df)

import requests
import pandas as pd

token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report'
parameter = {
    "securities_trader_id": "1020",
    "date": "2022-06-16",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
token = "" # Refer to login to obtain the API key
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report'
response = httr::GET(
    url = url,
    query = list(
        securities_trader_id="1020",
        start_date= "2022-06-16",
        token = token # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 合庫 122.25 19000 0 1020 0050 2022-06-16
1 合庫 122.3 80000 0 1020 0050 2022-06-16
2 合庫 122.35 10000 0 1020 0050 2022-06-16
3 合庫 122.5 1300 0 1020 0050 2022-06-16
4 合庫 122.55 20000 0 1020 0050 2022-06-16
... ... ... ... ... ... ... ...
3211 合庫 107 1000 50000 1020 9958 2022-06-16
3212 合庫 107.5 0 32000 1020 9958 2022-06-16
3213 合庫 108 0 2000 1020 9958 2022-06-16
3214 合庫 108.5 150 0 1020 9958 2022-06-16
3215 合庫 16.05 1000 0 1020 9962 2022-06-16 
{
    securities_trader: str, # securities trader name
    price: float64, # transaction price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Fetch all data for a specific date at once (available only to sponsorpro members)

(Due to the large data volume, each request only provides one day's data.)

  • Data range: 2021-06-30 ~ now.
  • Providing the dataset and date parameters returns the branch-level trading data of all stocks for that day.
  • Downloads the whole-day parquet via a signed URL, avoiding file-by-file queries — suitable for batch analysis covering the entire market.
  • Data update time: Monday to Friday 21:00. The actual update time is based on the API data.
  • Some data is missing on these dates: 2022-10-31~2022-11-03, 2023-01-11~2023-01-17.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_trading_daily_report(
    date='2022-06-16',
    use_object=True,
)

import io
import requests
import pandas as pd

url = "https://api.finmindtrade.com/api/v4/storage_objects"
token = "" # Refer to login to obtain the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockTradingDailyReport",
    "date": "2022-06-16",
}
resp = requests.get(url, headers=headers, params=parameter)
data = pd.read_parquet(io.BytesIO(resp.content))
print(data.head())

library(httr)
library(data.table)
library(dplyr)
library(arrow)

url = 'https://api.finmindtrade.com/api/v4/storage_objects'
token = "" # Refer to login to obtain the token
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockTradingDailyReport",
        date= "2022-06-16"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
con = content(response, "raw")
data <- read_parquet(con)
close(con)
head(data)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 合庫 508 4000 2000 1020 2330 2022-06-16
1 合庫 509 3480 0 1020 2330 2022-06-16
2 合庫 510 2310 50 1020 2330 2022-06-16
3 合庫 511 1169 0 1020 2330 2022-06-16
4 合庫 512 1300 10000 1020 2330 2022-06-16 
{
    securities_trader: str, # securities trader name
    price: float64, # deal price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Taiwan Stock Warrant Trading Daily Report by Branch (query by stock_id) TaiwanStockWarrantTradingDailyReport (only available for sponsor members)

  • Data range: 2023-06-21 ~ now
  • Due to the large data volume, only one day of data is provided per request.
  • Data update time Monday to Friday 01:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_warrant_trading_daily_report(
    stock_id="084655",
    date='2023-06-21',
)
print(df)

import requests
import pandas as pd

token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_warrant_trading_daily_report'
parameter = {
    "data_id": "084655",
    "date": "2023-06-21",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_warrant_trading_daily_report'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        data_id="084655",
        start_date= "2023-06-21",
        token = token # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

from FinMind.data import DataLoader
from loguru import logger
import datetime

api = DataLoader()
api.login_by_token(api_token='token')

date = '2025-12-08'
start = datetime.datetime.now()
df = api.taiwan_stock_warrant_trading_daily_report(
    date=date,
    use_async=True,
)
cost = datetime.datetime.now() - start
logger.info(cost)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 元富 2.48 0 4000 5920 084655 2023-06-21
1 凱基 2.48 4000 0 9200 084655 2023-06-21 
{
    securities_trader: str, # securities trader name
    price: float64, # transaction price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Taiwan Stock Warrant Trading Daily Report by Branch (query by securities_trader_id) TaiwanStockWarrantTradingDailyReport (only available for sponsor members)

  • Data range: 2023-06-21 ~ now
  • Due to the large data volume, only one day of data is provided per request.
  • Data update time Monday to Friday 23:00, the actual update time is based on the API data.

Example

import requests
import pandas as pd

url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_warrant_trading_daily_report'
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "securities_trader_id": "5920",
    "date": "2023-06-21",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_warrant_trading_daily_report'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        securities_trader_id="5920",
        start_date= "2023-06-21",
        token = token # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 元富 0.97 50000 0 5920 07741U 2023-06-21
1 元富 0.98 50000 0 5920 07741U 2023-06-21
2 元富 1.52 100000 0 5920 07742U 2023-06-21
3 元富 1.56 49000 0 5920 07742U 2023-06-21 
{
    securities_trader: str, # securities trader name
    price: float64, # transaction price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Fetch all warrant trading data for a specific date at once (available only to sponsorpro members)

(Due to the large data volume, each request only provides one day's data.)

  • Data range: 2023-06-21 ~ now.
  • Providing the dataset and date parameters returns the branch-level trading data of all warrants for that day.
  • Downloads the whole-day parquet via a signed URL, avoiding file-by-file queries — suitable for batch analysis covering all warrants in the market.
  • Updated daily. The actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_warrant_trading_daily_report(
    date='2023-06-21',
    use_object=True,
)

import io
import requests
import pandas as pd

url = "https://api.finmindtrade.com/api/v4/storage_objects"
token = "" # Refer to login to obtain the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockWarrantTradingDailyReport",
    "date": "2023-06-21",
}
resp = requests.get(url, headers=headers, params=parameter)
data = pd.read_parquet(io.BytesIO(resp.content))
print(data.head())

library(httr)
library(data.table)
library(dplyr)
library(arrow)

url = 'https://api.finmindtrade.com/api/v4/storage_objects'
token = "" # Refer to login to obtain the token
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockWarrantTradingDailyReport",
        date= "2023-06-21"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
con = content(response, "raw")
data <- read_parquet(con)
close(con)
head(data)

Output

securities_trader price buy sell securities_trader_id stock_id date
0 元富 2.48 0 4000 5920 084655 2023-06-21
1 凱基 2.48 4000 0 9200 084655 2023-06-21 
{
    securities_trader: str, # securities trader name
    price: float64, # deal price
    buy: int32, # shares bought
    sell: int32, # shares sold
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str # date
}

Taiwan Stock Government Bank Buy/Sell TaiwanStockGovernmentBankBuySell (only available for sponsor members)

  • Data range: 2021-06-30 ~ now
  • Due to the large data volume, only one day of data is provided per request.
  • Data update time Monday to Friday 23:30, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_government_bank_buy_sell(
    start_date='2023-01-17',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockGovernmentBankBuySell",
    "start_date": "2023-01-17",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockGovernmentBankBuySell",
        start_date= "2023-01-17"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id buy_amount sell_amount buy sell bank_name
0 2023-01-17 0050 43992298.6 53309904.25 372595 451744 兆豐
1 2023-01-17 5202 288.0 303.50 20 20 第一
2 2023-01-17 5202 0.0 59.45 0 4 華南
3 2023-01-17 5203 82800.0 0.00 1000 0 兆豐
4 2023-01-17 5203 249000.0 583600.00 3000 7000 臺銀 
{
    date: str, # date
    stock_id: str, # stock symbol
    buy_amount: float64, # buy amount
    sell_amount: float64, # sell amount
    buy: int64, # shares bought
    sell: int64, # shares sold
    bank_name: str # bank name
}

Taiwan Total Exchange Margin Maintenance TaiwanTotalExchangeMarginMaintenance (only available for backer, sponsor members)

  • Data range: 2001-01-05 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_total_exchange_margin_maintenance(
    start_date='2024-04-01',
    end_date='2024-05-01'
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanTotalExchangeMarginMaintenance",
    "start_date": "2020-04-01",
    "end_date": "2020-05-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanTotalExchangeMarginMaintenance",
        start_date= "2024-04-01",
        end_date='2024-05-01'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date TotalExchangeMarginMaintenance
0 2024-04-01 166.007
1 2024-04-02 167.079
2 2024-04-03 167.085
3 2024-04-08 167.119
4 2024-04-09 167.095 
{
    date: str, # date
    TotalExchangeMarginMaintenance: float64 # total exchange margin maintenance ratio
}

Daily Securities Trader Branch Aggregate Statistics TaiwanStockTradingDailyReportSecIdAgg (only available for sponsor members)

  • Provides Taiwan stock trading by branch information for listed (TWSE), OTC, and emerging stocks!
  • Data range: 2021-06-30 ~ now
  • Data update time Monday to Friday 21:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_trading_daily_report_secid_agg(
    stock_id="2330",
    securities_trader_id="1020",
    start_date= "2024-07-01",
    end_date="2024-07-15",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report_secid_agg"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "data_id": "2330",
    "securities_trader_id": "1020",
    "start_date": "2024-07-01",
    "end_date": '2024-07-15',
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/taiwan_stock_trading_daily_report_secid_agg'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        data_id="2330",
        securities_trader_id="1020",
        start_date= "2024-07-01",
        end_date='2024-07-15'
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

securities_trader securities_trader_id stock_id date buy_volume sell_volume buy_price sell_price
0 合庫 1020 2330 2024-07-01 12157 12460 968.08 973.84
0 合庫 1020 2330 2024-07-02 12735 21885 964.54 964.63
0 合庫 1020 2330 2024-07-03 10535 29381 973.16 974.69
0 合庫 1020 2330 2024-07-04 28107 59459 1001.99 1000.88
0 合庫 1020 2330 2024-07-05 10435 11075 1004.18 1004.5 
{
    securities_trader: str, # securities trader name
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str, # date
    buy_volume: int64, # total shares bought
    sell_volume: int64, # total shares sold
    buy_price: float, # average buy price
    sell_price: float, # average sell price
}

Block Trading Daily Report TaiwanStockBlockTradingDailyReport (only available for sponsor members)

  • Data range: 2026-04-28 ~ now
  • Only requires a date input, no stock symbol needed.

Example

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
parameter = {
    "dataset": "TaiwanStockBlockTradingDailyReport",
    "start_date": "2026-04-28",
    "token": token,
}
resp = requests.get(url, params=parameter)
data = resp.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
url = "https://api.finmindtrade.com/api/v4/data"
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockBlockTradingDailyReport",
        start_date="2026-04-28",
        token=""
    )
)
data = content(response)
df = do.call('rbind', lapply(data$data, as.data.frame))
head(df)

Output

{
    securities_trader: str, # securities trader
    price: float, # price
    buy: int, # shares bought
    sell: int, # shares sold
    trade_type: str, # trade type (paired/conversion, etc.)
    securities_trader_id: str, # securities trader code
    stock_id: str, # stock symbol
    date: str, # date
}

Block Trade Daily Transactions TaiwanStockBlockTrade (only available for sponsor members)

  • Data range: 2005-04-04 ~ now
  • Covers block trades (tick-by-tick) for listed (TWSE) and OTC stocks.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_block_trade(
    stock_id="2330",
    start_date="2026-04-01",
    end_date="2026-04-30",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockBlockTrade",
    "data_id": "2330",
    "start_date": "2026-04-01",
    "end_date": "2026-04-30",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockBlockTrade",
        data_id="2330",
        start_date="2026-04-01",
        end_date="2026-04-30",
        token=token
    )
)
data = content(response)
df = do.call("rbind", lapply(data$data, as.data.frame))
head(df)

Output

date stock_id trade_type price volume trading_money
0 2026-04-01 2330 配對交易 1871.29 25000 46782250
1 2026-04-01 2330 配對交易 1871.28 25000 46782000
2 2026-04-01 2330 配對交易 1865.28 15000 27979200
3 2026-04-01 2330 配對交易 1855 1078000 1999690000
4 2026-04-01 2330 配對交易 1849.42 137104 253562880 
{
    date: str, # date
    stock_id: str, # stock symbol
    trade_type: str, # trade type (paired trade, etc.)
    price: float, # transaction price
    volume: int, # transaction volume
    trading_money: int, # trading money
}

Loan Collateral Balance TaiwanStockLoanCollateralBalance (only available for sponsor members)

  • Data range: 2006-10-02 ~ now
  • Data update time Monday to Friday after market close, the actual update time is based on the API data.
  • Includes the collateral balance (in thousand shares) for margin purchase, securities firm securities-business loans, securities firm unrestricted-purpose loans, securities finance secured loans, and securities finance settlement margin for listed (TWSE) and OTC stocks.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_loan_collateral_balance(
    stock_id="2330",
    start_date="2026-04-01",
    end_date="2026-04-30",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockLoanCollateralBalance",
    "data_id": "2330",
    "start_date": "2026-04-01",
    "end_date": "2026-04-30",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockLoanCollateralBalance",
        data_id="2330",
        start_date="2026-04-01",
        end_date="2026-04-30",
        token=token
    )
)
data = content(response)
df = do.call("rbind", lapply(data$data, as.data.frame))
head(df)

Output

date stock_id market MarginPreviousDayBalance MarginBuy MarginSell MarginCashRedemption MarginCurrentDayBalance MarginNextDayQuota SecuritiesFirmLoanPreviousDayBalance SecuritiesFirmLoanBuy SecuritiesFirmLoanSell SecuritiesFirmLoanCashRedemption SecuritiesFirmLoanReplacement SecuritiesFirmLoanCurrentDayBalance SecuritiesFirmLoanNextDayQuota UnrestrictedLoanPreviousDayBalance UnrestrictedLoanBuy UnrestrictedLoanSell UnrestrictedLoanCashRedemption UnrestrictedLoanReplacement UnrestrictedLoanCurrentDayBalance UnrestrictedLoanNextDayQuota SecuritiesFinanceSecuredLoanPreviousDayBalance SecuritiesFinanceSecuredLoanBuy SecuritiesFinanceSecuredLoanSell SecuritiesFinanceSecuredLoanCashRedemption SecuritiesFinanceSecuredLoanReplacement SecuritiesFinanceSecuredLoanCurrentDayBalance SecuritiesFinanceSecuredLoanNextDayQuota SettlementMarginPreviousDayBalance SettlementMarginBuy SettlementMarginSell SettlementMarginCashRedemption SettlementMarginReplacement SettlementMarginCurrentDayBalance SettlementMarginNextDayQuota
0 2026-04-01 2330 集中市場 26608 1268 1485 38 26353 6483131 21 0 0 0 0 21 1296626 119953 851 142 46 215 120401 2593252 16700 111 11 30 6 16764 0 0 0 0 0 0 0 1296626
1 2026-04-02 2330 集中市場 26353 1312 873 194 26598 6483131 21 0 0 0 0 21 1296626 120401 621 139 70 416 120397 2593252 16764 134 6 19 18 16855 0 0 0 0 0 0 0 1296626
2 2026-04-07 2330 集中市場 26598 391 970 19 26000 6483131 21 0 0 0 0 21 1296626 120397 539 54 36 236 120610 2593252 16855 160 2 67 29 16917 0 0 0 0 0 0 0 1296626
3 2026-04-08 2330 集中市場 26000 2882 2057 28 26797 6483131 21 65 0 0 0 86 1296626 120616 776 75 134 222 120961 2593252 16917 88 21 28 9 16947 0 0 0 0 0 0 0 1296626
4 2026-04-09 2330 集中市場 26604 998 953 13 26636 6483131 86 0 65 0 0 21 1296626 120961 481 19 83 198 121142 2593252 16947 126 1 27 13 17032 0 0 0 0 0 0 0 1296626 
{
    date: str, # date
    stock_id: str, # stock symbol
    market: str, # market type (TWSE/OTC)
    MarginPreviousDayBalance: int, # margin purchase - previous day balance
    MarginBuy: int, # margin purchase - buy
    MarginSell: int, # margin purchase - sell
    MarginCashRedemption: int, # margin purchase - cash redemption
    MarginCurrentDayBalance: int, # margin purchase - current day balance
    MarginNextDayQuota: int, # margin purchase - next business day quota
    SecuritiesFirmLoanPreviousDayBalance: int, # securities firm securities-business loan - previous day balance
    SecuritiesFirmLoanBuy: int, # securities firm securities-business loan - buy
    SecuritiesFirmLoanSell: int, # securities firm securities-business loan - sell
    SecuritiesFirmLoanCashRedemption: int, # securities firm securities-business loan - cash redemption
    SecuritiesFirmLoanReplacement: int, # securities firm securities-business loan - replacement
    SecuritiesFirmLoanCurrentDayBalance: int, # securities firm securities-business loan - current day balance
    SecuritiesFirmLoanNextDayQuota: int, # securities firm securities-business loan - next business day quota
    UnrestrictedLoanPreviousDayBalance: int, # securities firm unrestricted-purpose loan - previous day balance
    UnrestrictedLoanBuy: int, # securities firm unrestricted-purpose loan - buy
    UnrestrictedLoanSell: int, # securities firm unrestricted-purpose loan - sell
    UnrestrictedLoanCashRedemption: int, # securities firm unrestricted-purpose loan - cash redemption
    UnrestrictedLoanReplacement: int, # securities firm unrestricted-purpose loan - replacement
    UnrestrictedLoanCurrentDayBalance: int, # securities firm unrestricted-purpose loan - current day balance
    UnrestrictedLoanNextDayQuota: int, # securities firm unrestricted-purpose loan - next business day quota
    SecuritiesFinanceSecuredLoanPreviousDayBalance: int, # securities finance secured loan - previous day balance
    SecuritiesFinanceSecuredLoanBuy: int, # securities finance secured loan - buy
    SecuritiesFinanceSecuredLoanSell: int, # securities finance secured loan - sell
    SecuritiesFinanceSecuredLoanCashRedemption: int, # securities finance secured loan - cash redemption
    SecuritiesFinanceSecuredLoanReplacement: int, # securities finance secured loan - replacement
    SecuritiesFinanceSecuredLoanCurrentDayBalance: int, # securities finance secured loan - current day balance
    SecuritiesFinanceSecuredLoanNextDayQuota: int, # securities finance secured loan - next business day quota
    SettlementMarginPreviousDayBalance: int, # securities finance settlement margin - previous day balance
    SettlementMarginBuy: int, # securities finance settlement margin - buy
    SettlementMarginSell: int, # securities finance settlement margin - sell
    SettlementMarginCashRedemption: int, # securities finance settlement margin - cash redemption
    SettlementMarginReplacement: int, # securities finance settlement margin - replacement
    SettlementMarginCurrentDayBalance: int, # securities finance settlement margin - current day balance
    SettlementMarginNextDayQuota: int, # securities finance settlement margin - next business day quota
}

Disposition Securities Period TaiwanStockDispositionSecuritiesPeriod (only available for backer, sponsor members)

  • Data range: 2001-01-01 ~ now
  • Data covers listed (TWSE), OTC, and emerging stocks.
  • Data update time Monday to Saturday 20:00~23:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_disposition_securities_period(
    stock_id="6477",
    start_date='2025-01-01',
    end_date='2025-02-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockDispositionSecuritiesPeriod",
    "data_id": "6477",
    "start_date": "2025-01-01",
    "end_date": "2025-02-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockDispositionSecuritiesPeriod",
        data_id= "6477",
        start_date= "2025-01-01",
        end_date= "2025-02-01"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table

Output

date stock_id stock_name disposition_cnt condition measure period_start period_end
0 2025-01-09 6477 安集 1 連續三次及當日沖銷標準 第一次處置 2025-01-10 2025-02-05 
{
    date: str, # announcement date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    disposition_cnt: int32, # cumulative count
    condition: str, # disposition condition
    measure: str, # disposition measure
    period_start: str, # disposition start date
    period_end: str, # disposition end date
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_disposition_securities_period(
    start_date='2025-01-09',
    end_date: "2025-01-09",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockDispositionSecuritiesPeriod",
    "start_date": "2025-01-09",
    "end_date": "2025-01-09",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockDispositionSecuritiesPeriod",
        start_date= "2025-01-09",
        end_date: "2025-01-09",
        token = "" # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id stock_name disposition_cnt condition measure period_start period_end
0 2025-01-09 6477 安集 1 連續三次及當日沖銷標準 第一次處置 2025-01-10 2025-02-05
1 2025-01-09 9103 美德醫療-DR 1 最近十個營業日已有六次 第二次處置 2025-01-10 2025-02-03 
{
    date: str, # announcement date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    disposition_cnt: int32, # cumulative count
    condition: str, # disposition condition
    measure: str, # disposition measure
    period_start: str, # disposition start date
    period_end: str, # disposition end date
}

Day Trading Borrowing Fee Rate TaiwanStockDayTradingBorrowingFeeRate (only available for backer, sponsor members)

  • Data range: 2015-06-01 ~ now
  • Data update time Monday to Friday 19:00~22:00, the actual update time is based on the API data.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_day_trading_borrowing_fee_rate(
    stock_id="2330",
    start_date='2024-12-01',
    end_date='2025-01-01',
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockDayTradingBorrowingFeeRate",
    "data_id": "2330",
    "start_date": "2024-12-01",
    "end_date": "2025-01-01",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockDayTradingBorrowingFeeRate",
        data_id= "2330",
        start_date= "2024-12-01",
        end_date= "2025-01-01"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table

Output

date stock_id stock_name InvestorBorrowedShares InvestorBorrowingFeeRate
0 2024-12-02 6477 安集 15000 7 
{
    date: str, # announcement date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    InvestorBorrowedShares: int32, # borrowed shares
    InvestorBorrowingFeeRate: float64, # borrowing fee rate
}

Get all data for a specific date in one request (only available for backer, sponsor members)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_disposition_securities_period(
    start_date='2024-12-02',
    end_date: "2024-12-02",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the API key
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockDayTradingBorrowingFeeRate",
    "start_date": "2024-12-02",
    "end_date": "2024-12-02",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data['data'])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the API key
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockDayTradingBorrowingFeeRate",
        start_date= "2024-12-02",
        end_date: "2024-12-02",
        token = "" # Refer to login to obtain the API key
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id stock_name InvestorBorrowedShares InvestorBorrowingFeeRate
0 2024-12-02 00631L 元大台灣50正2 5000 1
1 2024-12-02 1438 三地開發 1 3
2 2024-12-02 2312 金寶 1000 2
3 2024-12-02 2324 仁寶 1000 0.1
4 2024-12-02 2330 台積電 15000 7 
{
    date: str, # announcement date
    stock_id: str, # stock symbol
    stock_name: str, # stock name
    InvestorBorrowedShares: int32, # borrowed shares
    InvestorBorrowingFeeRate: float64, # borrowing fee rate
}

Taiwan Active ETF List TaiwanStockActiveETFInfo

  • This table lists Taiwan-listed active ETFs (TWSE-listed + TPEx OTC), including the ETF code, name, ETF category, and market type!
  • category is the ETF category: domestic (invests domestically) / foreign (invests cross-border)
  • type is the market type: twse (listed) / tpex (OTC)

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_active_etf_info()

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to login to obtain the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockActiveETFInfo",
}
resp = requests.get(url, headers=headers, params=parameter)
data = resp.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
library(dplyr)
url = 'https://api.finmindtrade.com/api/v4/data'
token = "" # Refer to login to obtain the token
response = httr::GET(
    url = url,
    query = list(
        dataset = "TaiwanStockActiveETFInfo"
    ),
    add_headers(Authorization = paste("Bearer", token))
)
data = content(response)
df = data$data %>%
do.call('rbind',.) %>%
data.table
head(df)

Output

date stock_id stock_name category type
0 2026-07-11 00980A 主動野村臺灣優選 domestic twse
1 2026-07-11 00981A 主動統一台股增長 domestic twse
2 2026-07-11 00982A 主動群益台灣強棒 domestic twse 
{
    date: str, # date
    stock_id: str, # ETF code
    stock_name: str, # ETF name
    category: str, # ETF category (domestic / foreign)
    type: str # market type (twse / tpex)
}

Active ETF Daily Holding TaiwanStockActiveETFHolding (only available for sponsor members)

  • Data range: 2025-05-05 ~ now
  • Data update time Monday to Saturday after market close, the actual update time is based on the API data.
  • Provides the full daily portfolio of Taiwan-listed active ETFs (TWSE + TPEx), including constituent code, name, asset type, shares, weight, market value and currency; buy/sell holdings can be derived by differencing consecutive trading days.
  • Query a single ETF via data_id (e.g. 00980A), or query all active ETF holdings of a given date by date only.
  • Note: the start date differs per ETF (depending on listing date and source history depth); some ETFs accumulate daily from launch.
  • Note: if an issuer does not publish an obtainable daily portfolio for a given active ETF, that ETF has no holding data for now (currently "主動貝萊德優投等 00985D") and will be backfilled once the source becomes available; such ETFs also have no corresponding TaiwanStockActiveETFHoldingChange.
  • Note: shares is an integer; market_value is disclosed per holding by only some active ETFs' daily portfolios — where it is not disclosed the field is 0 (you can estimate it as shares times the constituent's closing price).
  • Note: an active ETF's portfolio includes derivatives and cash/liability line items, not only stocks. For short derivative positions (written options, short futures) both shares (contracts) and market_value can be negative; liability line items (e.g. payables) can also have a negative market_value. Use asset_type (stock/bond/futures/option/cash/etf/repo/other) to filter — e.g. take asset_type == "stock" for equity holdings only.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_active_etf_holding(
    stock_id="00980A",
    start_date="2025-05-05",
    end_date="2025-05-31",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to the login to obtain the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockActiveETFHolding",
    "data_id": "00980A",
    "start_date": "2025-05-05",
    "end_date": "2025-05-31",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to the login to obtain the token
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockActiveETFHolding",
        data_id="00980A",
        start_date="2025-05-05",
        end_date="2025-05-31",
        token=token
    )
)
data = content(response)
df = do.call("rbind", lapply(data$data, as.data.frame))
head(df)

Output

date stock_id component_stock_id component_stock_name asset_type shares weight market_value currency
0 2025-05-05 00980A 2330 台灣積體電路製造 stock 634000 9.44 0 TWD
1 2025-05-05 00980A 2454 聯發科技 stock 269000 6.49 0 TWD
2 2025-05-05 00980A 2308 台達電子工業 stock 475000 5.41 0 TWD 
{
    date: str, # date (holding as-of date)
    stock_id: str, # ETF id (data_id)
    component_stock_id: str, # constituent code (TW code / foreign ticker / bond CUSIP)
    component_stock_name: str, # constituent name
    asset_type: str, # asset type (stock/bond/futures/option/cash/etf/repo/other)
    shares: int, # shares (par value for bonds, contracts for futures)
    weight: float, # weight (%)
    market_value: float, # market value
    currency: str, # currency
}

Active ETF Daily Holding Change TaiwanStockActiveETFHoldingChange (only available for sponsor members)

  • Data range: 2025-05-05 ~ now
  • Data update time Monday to Saturday after market close, the actual update time is based on the API data.
  • Derived from "Active ETF Daily Holding TaiwanStockActiveETFHolding": differencing the constituent shares of consecutive trading days yields which constituents each active ETF bought/sold that day.
  • buy is the shares bought that day (increase in constituent shares; 0 if none), sell is the shares sold that day (decrease in constituent shares, as a positive value; 0 if none); both are integers (no decimals), and exactly one of buy/sell is non-zero per row.
  • Query a single ETF via data_id (e.g. 00980A), or query all active ETF holding changes of a given date by date only.
  • Note: creations/redemptions scale constituent shares proportionally and are included in buy/sell; therefore buy/sell reflect the change in held shares and are not the manager's net discretionary buy/sell.

Example

from FinMind.data import DataLoader

api = DataLoader()
# api.login_by_token(api_token='token')
df = api.taiwan_stock_active_etf_holding_change(
    stock_id="00980A",
    start_date="2025-05-05",
    end_date="2025-05-31",
)

import requests
import pandas as pd
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to the login to obtain the token
headers = {"Authorization": f"Bearer {token}"}
parameter = {
    "dataset": "TaiwanStockActiveETFHoldingChange",
    "data_id": "00980A",
    "start_date": "2025-05-05",
    "end_date": "2025-05-31",
}
data = requests.get(url, headers=headers, params=parameter)
data = data.json()
data = pd.DataFrame(data["data"])
print(data.head())

library(httr)
library(data.table)
url = "https://api.finmindtrade.com/api/v4/data"
token = "" # Refer to the login to obtain the token
response = httr::GET(
    url = url,
    query = list(
        dataset="TaiwanStockActiveETFHoldingChange",
        data_id="00980A",
        start_date="2025-05-05",
        end_date="2025-05-31",
        token=token
    )
)
data = content(response)
df = do.call("rbind", lapply(data$data, as.data.frame))
head(df)

Output

date stock_id component_stock_id component_stock_name buy sell
0 2025-05-06 00980A 2330 台灣積體電路製造 12000 0
1 2025-05-06 00980A 2454 聯發科技 0 5000
2 2025-05-06 00980A 2308 台達電子工業 8000 0 
{
    date: str, # date
    stock_id: str, # ETF id (data_id)
    component_stock_id: str, # constituent code
    component_stock_name: str, # constituent name
    buy: int, # shares bought that day (increase in constituent shares; 0 if none)
    sell: int, # shares sold that day (decrease in constituent shares, positive; 0 if none)
}