convert dataframe column from list of strings to tuples

user_8643323 0
Viewed 2402

I have urls in a list. This is an element of a dataframe. I need convert each of these lists of strings to a hashable type like tuples. I've read that tuple(a,) with the comma, preserves the strings inside the lists when converting. I can't seem to get this to work when applying to a whole column of dataframe. prob missing something simple

df['url'] = tuple(df['url',]) ...doesn't work

flatframe['url'] = flatframe['url'].apply(tuple) ...works but doesn't preserve strings

here are a couple of rows of data:

index artist ranking song songurl songtext artisturl year

2280    (Lady Antebellum,)  81  [Bartender (Lady Antebellum song)]  [/wiki/Bartender_(Lady_Antebellum_song)]    "Bartender (Lady Antebellum song)"  /wiki/Lady_Antebellum   2014

2281    (Naughty Boy, Sam Smith)    82  [La La La (Naughty Boy song)]   [/wiki/La_La_La_(Naughty_Boy_song)] "La La La (Naughty Boy song)"   [/wiki/Naughty_Boy, /wiki/Sam_Smith_(singer)]   2014

2282    (Robin Thicke, T.I., Pharrell Williams) 83  [Blurred Lines] [/wiki/Blurred_Lines]   "Blurred Lines" [/wiki/Robin_Thicke, /wiki/T.I., /wiki/Pharrel...   2014

2283    (Lady Gaga, R. Kelly)   84  [Do What U Want]    [/wiki/Do_What_U_Want]  "Do What U Want"    [/wiki/Lady_Gaga, /wiki/R._Kelly]   2014
pandas dataframe tuples python
1 Answers

Lets say your dataframe is something like this:

import pandas as pd
pd.set_printoptions(max_columns=10)
df = pd.DataFrame(
[[2280, ("Lady Antebellum"),  81,  ["Bartender (Lady Antebellum song)"],  ["/wiki/Bartender_(Lady_Antebellum_song)"],    "Bartender (Lady Antebellum song)",  "/wiki/Lady_Antebellum",  2014],
[2281, "(Naughty Boy, Sam Smith)",    82,  ["La La La (Naughty Boy song)"],   ["/wiki/La_La_La_(Naughty_Boy_song)"], "La La La (Naughty Boy song)",   ["/wiki/Naughty_Boy", "/wiki/Sam_Smith_(singer)"],   2014],
[2282, "(Robin Thicke, T.I., Pharrell Williams)", 83,  ["Blurred Lines"], ["/wiki/Blurred_Lines"],   "Blurred Lines", ["/wiki/Robin_Thicke", "/wiki/T.I. /wiki/Pharrel"],   2014],
[2283, "(Lady Gaga, R. Kelly)",   84,  ["Do What U Want"],    ["/wiki/Do_What_U_Want"],  "Do What U Want",    ["/wiki/Lady_Gaga", "/wiki/R._Kelly"],   2014]],
columns = ["index", "artist", "ranking", "song", "songurl", "songtext", "artisturl", "year"])

Then you can try with:

df.artisturl = df.artisturl.apply(lambda x: tuple(x) if type(x)!= str else tuple([x]))

This will apply tuple only to entries that are not strings, and convert to list and then to tuple entries that are strings. As if it is a string and you apply tuple it will give a tuple with each character as entries.

Your column artisturl would then look as:

>>> df.artisturl
0                           ('/wiki/Lady_Antebellum',)
1    ('/wiki/Naughty_Boy', '/wiki/Sam_Smith_(singer)')
2    ('/wiki/Robin_Thicke', '/wiki/T.I. /wiki/Pharr...
3                ('/wiki/Lady_Gaga', '/wiki/R._Kelly')
Name: artisturl
user_53992680
Related