Technolohic27 Technolohic27 - 5 months ago 13
Python Question

How to delete entire row of data set given a condition on a column in csv file?

Here is a snippet of the following data-set in csv format:

quantity revenue time_x transaction_id user_id
1 0 57:57.0 0 0 0
1 0 18:59.0 0 1


I want to delete the entire row when the user_id is empty. How do I do this in python? So far, here's my code:

activity = pd.read_csv("activity(delimited).csv", delimiter=';', error_bad_lines=False, dtype=object)
impression = pd.read_csv("impression(delimited).csv", delimiter=';', error_bad_lines=False, dtype=object)
click = pd.read_csv("click(delimited).csv", delimiter=';', error_bad_lines=False, dtype=object)

pre_merge = activity.merge(impression, on="user_id", how="outer")
merged = pre_merge.merge(click, on="user_id", how="outer")
merged.to_csv("merged.csv", index=False)

open_merged = pd.read_csv("merged.csv", delimiter=',', error_bad_lines= False, dtype=object)

filtered_merged = open_merged.dropna(axis='columns', how='all')


Also, how can I write the code in an efficient manner?

Answer

With Pandas:

import pandas as pd

df = pd.read_csv("path/to/csv/data.csv", delimiter=';', error_bad_lines=False)
df = df[pd.notnull(df.user)] # boolean indexing

# Shift user_id to first column
df = df.set_index("user_id")
df = df.reset_index()

df.to_csv("path/to/csv/data.csv", index=False)

The bracket notation allows you provide an iterable of boolean values. This is called boolean indexing. Similar concepts and syntax are used in numpy, matlab and R