Converting a pandas series to a string array

Question

I have a pandas dataframe:

ind  
0   ['C']
1   ['C']
2   ['C']
3   ['C']
4   ['E']
5   ['E']

I want to convert it into a string: CCCCEE

possible duplicate: https://stackoverflow.com/questions/41400381/python-pandas-concatenate-a-series-of-strings-into-one-string — Dodge, May 22 '18 at 22:16
Possible duplicate of [Python Pandas concatenate a Series of strings into one string](https://stackoverflow.com/questions/41400381/python-pandas-concatenate-a-series-of-strings-into-one-string) — Dodge, May 22 '18 at 22:17

score 4 · Answer 1 · answered May 22 '18 at 21:47

4

You can using str

df['ind'].str[0].sum()
Out[197]: 'CCCCEE'

answered May 22 '18 at 21:47

BENY

317,841
20
164
234

ind is not an actual column name, just the default python dataframe indexing. Could you please suggest how should I proceed? – kdba May 22 '18 at 21:55
@KrishnaAnapindi You mean: `df.index.str[0].sum()`? – Anton vBR May 22 '18 at 21:57
gives an error saying 'Index' object has no attribute to sum – kdba May 22 '18 at 21:59
@KrishnaAnapindi it is not like a index in your sample data – BENY May 22 '18 at 21:59
Yeah just replace 'ind' with the name of the column containing the strings – Dodge May 22 '18 at 22:00
@KrishnaAnapindi seems like it is Series not data frame , `df.str[0].sum()` – BENY May 22 '18 at 22:01
@Wen If it is a series however OP should avoid to use the variable name `df` and change it to `s` which is the common variable name for series. Good guess though, probably it. – Anton vBR May 22 '18 at 22:10
@AntonvBR my guess that ind is index name ：-） – BENY May 22 '18 at 22:10

jpp · Answer 2 · 2018-05-22T22:15:42.857

3

Using itertools.chain:

from itertools import chain

df = pd.DataFrame({'ind': [['C'], ['C'], ['C'], ['C'], ['E'], ['E']]})

res = ''.join(chain.from_iterable(df['ind']))

print(res)

CCCCEE

edited May 22 '18 at 22:15

answered May 22 '18 at 21:47

jpp

159,742
34
281
339

Sure but shouldn't `''.join(chain.from_iterable(df['ind']))` suffice? – Anton vBR May 22 '18 at 21:55
1

@AntonvBR, The `list` conversion is intentional, [see here why](https://stackoverflow.com/a/37782238/9209546). You're right that accessing `values` is unnecessary, updated. – jpp May 22 '18 at 21:59
Oh.. it is that same old thing. Yeah that is probably right but I feel it looses readability when applying multiple layers of parenthesis. – Anton vBR May 22 '18 at 22:03
Ok I ran some timings with a very large dataframe and I interpret it like this: the `str.join` will convert a generator to a list before applying the `join` and it is therefore recommended to do `[]` a list comprehension over a `()` expression. However when the returning value as in this case is an iterator there is no point to convert it to a list as this will be done by the `str.join` anyway. I therefore think you are wrong in this particular case and that the `list()` only creates the extra layer of parenthesis. – Anton vBR May 22 '18 at 22:08
@AntonvBR, That's interesting, and I agree `list` is expensive. Looks like I need to dig a little deeper. Thanks :). – jpp May 22 '18 at 22:16
In any case. Nice solution. Sorry for spamming comments. – Anton vBR May 22 '18 at 22:17
@AntonvBR, No.. in fact, I think this is precisely the purpose of comments. Learn something new! – jpp May 22 '18 at 22:17
OK last thing. Let's leave this question open for the future. The answer might not be right but I strongly believe it is. – Anton vBR May 22 '18 at 22:18

score 0 · Answer 3 · answered May 22 '18 at 21:51

0

You can do this:

chain=""
for index, row in df.iterrows():
    chain=chain+row['column']
return chain

If you have a problem iterating over a dataframe, you can check this How to iterate over rows in a DataFrame in Pandas?

answered May 22 '18 at 21:51

sebashc3712

65
7

Converting a pandas series to a string array

3 Answers3