CODEWITHSUNDEEP
pandas
RJF
RJF

Reputation: 447

Replace rows in one column with values from another if a condition met

I have the following dataframe:

df = pd.DataFrame([[1, 2, "Yes"], [4, 5, "No"], [7, 8, "Yes"], [3, 5, "No"]],
...      index=['a', 'b', 'c', 'd' ],
...      columns=['column_1', 'column_2', 'condition'])

    column_1    column_2    condition
a       1           2        Yes
b       4           5        No
c       7           8        Yes
d       3           5        No

I would like to replace the row values in column_1 with corresponding values in column_2 when the condition is "Yes". I have been trying to do this using the following code:

df.loc[df['condition'] == "Yes", ['column_1']] = df.ix[df['column_1']]

Any hint is much appreciated!

Upvotes: 1

Views: 183

Answers (1)

cs95
cs95

Reputation: 403128

ix is deprecated. But aside from that, you do not need any slicer/selector here. Just select with loc on the LHS, and pandas will perform index-aligned assignment. So, this is enough:

df.loc[df['condition'] == "Yes", 'column_2'] = df['column_1']

In indexed aligned assignment, only the indices that are required fromdf['column_1'] are pulled in. From,

df.loc[df['condition'] == "Yes", 'column_2'] 

a    1
c    7
Name: column_2, dtype: int64

We see only the values for index "a" and "c" are needed from df['column_1'], everything else is ignored.

Upvotes: 1

Related Questions