CODEWITHSUNDEEP
pythonpandas
Tjorriemorrie
Tjorriemorrie

Reputation: 17282

How to set row value in pandas dataframe?

I need to set the first row's value of a column for every group to False.

start_ix = group.head(1).index

print(start_ix)

Int64Index([20], dtype='int64')

print df.iloc[start_ix]['start']

20 False

Name: start, dtype: bool

df.iloc[start_ix]['start'] = True

But the value stays False upon inspecting the DataFrame.

Upvotes: 3

Views: 14018

Answers (1)

EdChum
EdChum

Reputation: 394041

So basically what you tried is what is called chained indexing and may or may not work:

df.iloc[start_ix]['start'] = True

It may help to think of this as chaining function calls, the problem here is that sometimes it returns a view which is what you intended, but other times it returns a copy resulting in no change in the original df.

So you should use ix which allows you to be explicit in indexing by integer value (or label) and then specifying the column label you want to set the data:

df.ix[start_ix,'start'] = True

The docs detail the recommended methods, normally a warning will be raised when you are setting the value of a potential copy but it cannot catch all instances.

Update

Since version 0.20.0 ix is deprecated in favour of loc so the above becomes

df.loc[start_ix,'start'] = True

Upvotes: 3

Related Questions