Pandas - 1

Import Pandas:

import pandas as pd

General settings

Command	Description
pd.set_option('display.max_columns', 1000)	Increase the number of columns displayed
pd.set_option('display.max_rows', 1000)	Increase the number of rows displayed
pd.set_option('display.max_colwidth', None)	Increase the cell size limit

Date operations

First, transform the column to datetime format specifying the string format to be parsed:

df[col] = pd.to_datetime(df[col], format='%Y-%m-%d', errors=_errors)

By setting _errors to raise (default), inconsistent dates will raise exceptions. To ignore them and set as missing use coerce.

Add days to a date column

df[col] = df[col] + pd.Timedelta(ndays, unit='D')

Add months to a date

pd.to_datetime(a_date_string, format='%Y-%m-%d') + pd.DateOffset(months=nmonths)

Get the time difference among two dates

(t2 - t1).days

In weeks

data['weeks_since_start_date'] = ledger_data.apply(
    lambda df:
        int(np.floor(
            (
                pd.to_datetime(the_date) -
                pd.to_datetime(df['start_date'])
            )/np.timedelta64(1, 'W')
        )), 
        axis=1
)

Remove timezone

Count number of rows and columns

The number of rows in a dataframe df is given by:

numRows = df.shape[0]
print(numRows)

The number of columns:

numCols = df.shape[1]
print(numCols)

---

Continue to Python-pandas (2)

Return to main page