pandas column names to list

Question

According to this thread: SO: Column names to list

It should be straightforward to do convert the column names to a list. But if i do:

df.columns.tolist()

I do get:

[u'q_igg', u'q_hcp', u'c_igg', u'c_hcp']

I know, i could get rid of the u and the ' . But i would like to just get the clean names as list without any hack around. Is that possible ?

This is correct, it just indicates that the strings are Unicode strings. — Simeon Visser, Nov 25 '14 at 14:23

score 23 · Accepted Answer · edited Jan 18 '23 at 12:51

23

Or, you could try:

df2 = df.columns.get_values()

which will give you:

array(['q_igg', 'q_hcp', 'c_igg', 'c_hcp'], dtype=object)

then:

df2.columns.tolist()

which gives you:

['q_igg', 'q_hcp', 'c_igg']

edited Jan 18 '23 at 12:51

Pav K.

2,548
2
19
29

answered Jan 23 '17 at 04:59

gincard

1,814
3
16
24

pretty verbose .. but maybe that's the only way ..? – WestCoastProjects Mar 11 '18 at 20:58
4

Slightly less verbose: `df.columns.values.tolist()` – snark Jun 12 '18 at 15:02
1

The `get_values()` method is depreciated: "FutureWarning: The 'get_values' method is deprecated and will be removed in a future version. Use '.to_numpy()' or '.array' instead." – Paroofkey Sep 02 '19 at 10:41
Please update your answer, since it is still the accepted answer. – Nyxynyx May 04 '20 at 17:31
try this : `list(df2)` – Omkar Darves Jan 14 '22 at 12:01

score 4 · Answer 2 · edited Jun 20 '20 at 09:12

4

Simple and easy way: df-dataframe variable name

df.columns.to_list()

this will give the list of the all columns name.

edited Jun 20 '20 at 09:12

Community

1
1

answered Sep 19 '19 at 11:15

brijesh_patel

41
2

score 3 · Answer 3 · answered Nov 25 '14 at 14:25

The list [u'q_igg', u'q_hcp', u'c_igg', u'c_hcp'] contains Unicode strings: the u indicates that they're Unicode strings and the ' are enclosed around each string. You can now use these names in any way you'd like in your code. See Unicode HOWTO for more details on Unicode strings in Python 2.x.

score 1 · Answer 4 · answered Nov 25 '14 at 14:29

1

If you're just interested in printing the name without an quotes or unicode indicators, you could do something like this:

In [19]: print "[" + ", ".join(df) + "]"
[q_igg, q_hcp, c_igg, c_hcp]

answered Nov 25 '14 at 14:29

chrisb

49,833
8
70
70

score 1 · Answer 5 · answered Nov 26 '14 at 07:40

As already mentioned the u means that its unicode converted. Anyway, the cleanest way would be to convert the colnames to ascii or something like that.

In [4]: cols
Out[4]: [u'q_igg', u'q_hcp', u'c_igg', u'c_hcp']

In [5]: [i.encode('ascii', 'ignore') for i in cols]
Out[5]: ['q_igg', 'q_hcp', 'c_igg', 'c_hcp'

The problem here is that you would lose special characters that are not encode in ascii.

A much more dirty solution would be to fetch the string representation of the list object and just replace the u. I would not use that but it might befit your needs in this special case ;-)

In [7]: repr(cols)
Out[7]: "[u'q_igg', u'q_hcp', u'c_igg', u'c_hcp']"
In [11]: x.replace("u", "")
Out[11]: "['q_igg', 'q_hcp', 'c_igg', 'c_hcp']"

see: https://docs.python.org/2/library/repr.html

Commenting on behalf of @AsheKetchum who doesn't have enough rep: The downside of `.replace` is that it might replace '**u**' if your original variables have u in their names. e.g. `"u'q_ugg'"` would become `"'q_gg'"` — Cory Klein, Feb 16 '17 at 20:52

score 0 · Answer 6 · answered Mar 22 '21 at 08:13

0

this will do the job

list(df2)

answered Mar 22 '21 at 08:13

Omkar Darves

164
1
5

pandas column names to list

6 Answers6