Fill missing rows based on index number gaps, why does it work? - pandas series

Question

So say i have a pandas series, as:

s = pd.Series([1,2],index=[0,3])

---

0    1
3    2
dtype: int64

And there's a gap between 0 and 3 in the index, so what i want is to add more rows to fill up the gaps to get the index of [0, 1, 2, 3].

So desired output would look like:

0    1.0
1    NaN
2    NaN
3    2.0
dtype: float64

And i did:

print(s.reindex(range(s.index.min(),s.index.max()+1)))

And it worked!

But why?

I expected a result of:

0    1.0
1    2.0
2    NaN
3    NaN
dtype: float64

But it doesn't, and gives expected one!

(you know, i was ready to create a question about how to do this, but while ready to show an attempt, i solved it :D, so asked a question why did it work :-) , lol )

Answer 1

The reason is simply because how reindex() is implemented.

If you take a look at the example given in the documentation, executing reindex() only adds the missing index in the specified range with NaN value. It does not suppose to change the index of the available entry.

Answer 2

Intrinsic data alignment. Basically, your source data is aligned with index 0 and 3. When you use reindex, you are creating new rows 1, and 2 and reusing 0 and 3.

Watch what happens if you do:

s.reindex([0,0,3,3])

Output:

0    1
0    1
3    2
3    2
dtype: int64

Pandas automatically using index alignment.

Or

s.reindex([1,2,5,6])

Output:

1   NaN
2   NaN
5   NaN
6   NaN
dtype: float64