CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
Sri Test
Sri Test

Reputation: 387

Dataset from original

I have a dataframe like this:

  a|b|c|d
1  true|false|true|false
2  false|true|true|false
3  true|true|true|false

I want this to be structured as :

a,c
b,c
a,b,c

and exported to a text file. How can we do that? I tried and could do the opposite but not this.Could you give me some idea?

Upvotes: 0

Views: 49

Answers (2)

ALollz
ALollz

Reputation: 59579

You could modify the column names, use dot (As True == 1 and False == 0), and then get rid of the trailing comma.

#df = df.replace({'true': True, 'false': False}) # If strings, not Bools

df.columns = [f'{x},' for x in df.columns]
df.dot(df.columns).str.strip(',')

#0      a,c
#1      b,c
#2    a,b,c
#dtype: object

Or more plainly, string join the columns after slicing by the Boolean row Series:

pd.Series([','.join(df.columns[x]) for _,x in df.iterrows()],
          index=df.index)

#0      a,c
#1      b,c
#2    a,b,c
#dtype: object

Upvotes: 3

Alexander
Alexander

Reputation: 109666

First, get the column names by splitting the first column on the | symbol. Then split the values in that column on the same symbol, and use a conditional comprehension to select the corresponding columns where the split value is true. Save the result back to a .csv file without the index.

cols = df.columns[0].split('|')
(df
 .iloc[:, 0].str
 .split('|')
 .apply(lambda bools: ','.join(col for col, b in zip(cols, bools) if b == 'true'))
 .to_csv(filename, index=False)
)

Upvotes: 1

Related Questions