pandas · Lesson 1.3

Reading CSV files

The read_csv options you'll use most: dates, chosen columns, and values that mean "missing".

Practice data: PayPal files

Made up for teaching, not real PayPal data. Read a file with pd.read_csv("accounts.csv"). Disclaimer

accounts.csv (300 rows): account_id, created_at, country, us_state, account_type

fx_rates.csv (260 rows): rate_date, currency, usd_rate

merchants.csv (60 rows): merchant_id, merchant_name, category, country

raw_payouts.csv (120 rows): payout_id, merchant_id, amount_text, paid_on

refunds.csv (179 rows): refund_id, txn_id, refund_date, amount

transactions.csv (2,980 rows): txn_id, account_id, merchant_id, txn_ts, amount, currency, status, risk_score, us_state

wallet_entries.csv (1,375 rows): entry_id, account_id, entry_date, amount

1Dates come in as text

read_csv reads date columns as plain text (str). parse_dates=["col"] turns them into real datetimes you can do date maths with.

import pandas as pd

txns = pd.read_csv("transactions.csv", parse_dates=["txn_ts"])
print(txns["txn_ts"].dtype)
print(txns["txn_ts"].min(), txns["txn_ts"].max())

PayPal data

Practice

Read transactions.csv with txn_ts as a datetime and store the number of days between the first and the last transaction in result.

import pandas as pd

txns = pd.read_csv("transactions.csv")
result = ...

PayPal data

2Only the columns you need

usecols=[...] reads just some columns, which keeps big files fast and small.

import pandas as pd

small = pd.read_csv("transactions.csv", usecols=["txn_id", "amount", "currency"])
small.head()

PayPal data

Practice

Read only the currency and amount columns of transactions.csv, and store the DataFrame in result.

import pandas as pd

result = ...

PayPal data

3Values that mean "missing"

pandas already treats empty cells and text like N/A as missing (NaN). na_values=[...] adds your own placeholders, like pending.

import pandas as pd

raw = pd.read_csv("raw_payouts.csv")
print(raw["amount_text"].isna().sum())
clean = pd.read_csv("raw_payouts.csv", na_values=["pending"])
print(clean["amount_text"].isna().sum())

PayPal data

Practice

Read raw_payouts.csv so that both N/A and pending count as missing, and store the number of missing amount_text values in result.

import pandas as pd

payouts = pd.read_csv("raw_payouts.csv")
result = ...

PayPal data

4Peek at a big file

nrows=n reads just the first n rows, a quick way to look at a file before loading all of it.

import pandas as pd

preview = pd.read_csv("transactions.csv", nrows=3)
preview

PayPal data

Practice

Read the first 10 rows of accounts.csv and store them in result.

import pandas as pd

result = ...

PayPal data

Tip

In real projects files usually come from a path like pd.read_csv("data/2026/orders.csv") or a URL. The options are the same.