Split by character quantity, create new columns with substrings Python

Question

I found this example on another post, but its for R, I need it for Python in a pandas dataframe

Original post = Split string every n characters new column

Suppose I have a data frame like this with a string vector, var2

var1	var2
1	abcdefghi
2	abcdefghijklmnop
3	abc
4	abcdefghijklmnopqrst

What is the most efficient way to split var2 every n characters into new columns until the end of each string,

e.g if every 4 characters, the output would like look like this:

var1	var2	new_var1	new_var2	new_var3	new_var4	new_var5
1	abcdefghi	abcd	efgh	i
2	abcdefghijklmnop	abcd	efgh	ijkl	mnop
3	abc	abc
4	abcdefghijklmnopqrst	abcd	efgh	ijkl	mnop	qrst

To make it more difficult, I don't know how big is the longest string in column, but I need the total column to be as many as needed based on the resulting N columns

Corralien · Accepted Answer · 2022-02-14T22:55:49.357

2

You can use:

N = 4

# Custom function to split string
split_string = lambda x: pd.Series([x[i:i+N] for i in range(0, len(x), N)])

new_var = df['var2'].apply(split_string).fillna('')
new_var.columns = 'new_var' + (new_var.columns + 1).astype(str).str.zfill(3)

df = df.join(new_var)

Output:

var1	var2	new_var001	new_var002	new_var003	new_var004	new_var005
1	abcdefghi	abcd	efgh	i
2	abcdefghijklmnop	abcd	efgh	ijkl	mnop
3	abc	abc
4	abcdefghijklmnopqrst	abcd	efgh	ijkl	mnop	qrst

edited Feb 14 '22 at 22:55

answered Feb 14 '22 at 21:46

Corralien

109,409
8
28
52

Thank you, it works fine. But why do you use ":i+4]" and later ", N)]". They should be both the same number right? – Random Person Feb 14 '22 at 22:04
Just one extra thing, can you add leading zeroes to the (new_var.columns + 1) – Random Person Feb 14 '22 at 22:38
001, 002.....010, 011 etc – Random Person Feb 14 '22 at 22:38
@RandomPerson. I updated my answer according to your comment. Can you check it please? – Corralien Feb 14 '22 at 22:56
You're welcome. Glad this help you. – Corralien Apr 17 '23 at 20:24

score 0 · Answer 2 · answered Feb 14 '22 at 22:27

One option is to do all the preprocesssing within vanilla python, before returning to Pandas (it should be fast/efficient):

outcome = [[ent[n:n+4] for n in range(0, len(ent), 4)] for ent in df.var2]
outcome = pd.DataFrame(outcome)
outcome.columns = outcome.columns.map(lambda col: f"new_var{col+1}")
pd.concat([df, outcome], axis="columns")

   var1                  var2 new_var1 new_var2 new_var3 new_var4 new_var5
0     1             abcdefghi     abcd     efgh        i     None     None
1     2      abcdefghijklmnop     abcd     efgh     ijkl     mnop     None
2     3                   abc      abc     None     None     None     None
3     4  abcdefghijklmnopqrst     abcd     efgh     ijkl     mnop     qrst

Split by character quantity, create new columns with substrings Python

2 Answers2