熊猫read_csv fillna

Question

I have some data that I am reading from a CSV file and one data frame column is recorded on a different time stamp interval (time series data) and I cant get a df.fillna(method = 'ffill').fillna(method = 'bfill') to work.

If I don't read the CSV file with a keep_default_na=False Python fills the gaps with a NaN but I would like the gaps to be blank so I can use the df.fillna(method = 'ffill')

import pandas as pd
import numpy as np

#read CSV file
df_raw = pd.read_csv('C:\\desktop\\combinedSP.csv', index_col='Date', parse_dates=True, keep_default_na=False)

df_raw.head()

df_raw2 = df_raw.fillna(method = 'ffill').fillna(method = 'bfill')

df_raw2.head()

It seems like no matter what I attempt I am not fixing the issue on the column labeled OAT :(

Any tips greatly appreciated, I have the data CSV file here loaded into my GitHub account.

Answer 1

When you do keep_default_na=False this means that what read_csv usually would read and parse to NaN it will no longer :

By default the following values are interpreted as NaN: '', '#N/A', '#N/AN/A', '#NA', '-1.#IND', '-1.#QNAN', '-NaN', '-nan', '1.#IND', '1.#QNAN', 'N/A', 'NA', 'NULL', 'NaN', 'n/a', 'nan', 'null' .

In this case, it's not parsing the empty string '' as NaN, it's keeping them as the empty string.

Drop that kwarg and the fillnas ought to work.