how to convert pandas series to tuple of index and value

Question

I'm looking for an efficient way to convert a series to a tuple of its index with its values.

s = pd.Series([1, 2, 3], ['a', 'b', 'c'])

I want an array, list, series, some iterable:

[(1, 'a'), (2, 'b'), (3, 'c')]

That `pd.Series()` call actually makes `['a', 'b', 'c']` the index and `[1, 2, 3]` the data, not what you might expect. But your example suggests you want 1,2,3 first hence to be the index. `list(s.items())` gives `[('a', 1), ('b', 2), ('c', 3)]`. So it seems to be what you need; maybe you just need to flip index and data in the declaration. — smci, Dec 10 '20 at 01:10

Divakar · Accepted Answer · 2018-08-04T05:57:13.690

63

Well it seems simply zip(s,s.index) works too!

For Python-3.x, we need to wrap it with list -

list(zip(s,s.index))

To get a tuple of tuples, use tuple() : tuple(zip(s,s.index)).

Sample run -

In [8]: s
Out[8]: 
a    1
b    2
c    3
dtype: int64

In [9]: list(zip(s,s.index))
Out[9]: [(1, 'a'), (2, 'b'), (3, 'c')]

In [10]: tuple(zip(s,s.index))
Out[10]: ((1, 'a'), (2, 'b'), (3, 'c'))

edited Aug 04 '18 at 05:57

answered Jul 19 '16 at 21:18

Divakar

218,885
19
262
358

4

should be `list(zip(s,s.index))` otherwise you get a zip object in python3 – Claudiu Creanga Jan 26 '18 at 12:25
this returns a zip object he is asking for a `tuple` – Seymour Apr 14 '18 at 23:59
1

@Seymour At the time of posting, I had access to Python-2, so missed out then. Edited to fix it for python3 scenario. – Divakar Apr 15 '18 at 06:10

smci · Answer 2 · 2022-01-14T10:23:55.027

31

s.items() or s.iteritems() do this.

(If you want to get the output as a list rather than an iterator, do: list(s.items()))

edited Jan 14 '22 at 10:23

answered Feb 16 '18 at 06:29

smci

32,567
20
113
146

abeboparebop · Answer 3 · 2016-07-20T11:39:35.107

One possibility is to swap the order of the index elements and the values from iteritems:

res = [(val, idx) for idx, val in s.iteritems()]

EDIT: @Divakar's answer is faster by about a factor of 2. Building a series of random strings for testing:

N = 100000
str_len = 4
ints = range(N)
strs = [None]*N
for i in ints:
    strs[i] = ''.join(random.choice(string.ascii_letters) for _ in range(str_len))
s = pd.Series(ints, strs)

Timings:

%timeit res = zip(s,s.index)
>>> 100 loops, best of 3: 14.8 ms per loop

%timeit res = [(val, idx) for idx, val in s.iteritems()]
>>> 10 loops, best of 3: 26.7 ms per loop

how to convert pandas series to tuple of index and value

3 Answers3