Using path-strings as index in pandas

Question

I am trying to create a dataframe with filepaths as index:

import os
import pandas as pd
pathnames = []
for i in range(5):
    pathnames.append(os.path.join('a',str(i)))

print(pathnames)
df = pd.DataFrame(index = pathnames)
df[pathnames[0]]

When using this example I get a key-error as pathnames[0] is interpreted as a\\0 (as when printed) but not as a\0. I expect the problem to be hidden in the raw string formatting. Can someone explain and help? pandas version tested: 0.24.2 and 1.1.0

`df[pathnames[0]]` selects columns, not indices. if you pass `df.loc[pathnames[0]]` you get a result — sammywemmy, Aug 19 '20 at 11:15
Thanks a lot. I guessed it was something trivial but I did not think the right way — user14130879, Aug 19 '20 at 12:36

score 0 · Answer 1 · answered Aug 19 '20 at 11:16

0

Use loc or iloc

import os
import pandas as pd
pathnames = []
for i in range(5):
    pathnames.append(os.path.join('a',str(i)))

print(pathnames)
df = pd.DataFrame(index = pathnames)
print(df.loc['a/0'])

answered Aug 19 '20 at 11:16

Krishna Singhal

631
6
9

Using path-strings as index in pandas

1 Answers1