Pandas: How to calculate number of changes considering previous record

Question

I have a dataframe like below:

import pandas as pd 
df = pd.DataFrame({"id":[1,1,1,1,1,1,2,2,2,2,3,3,3,3], 
                  "val":[0,1,1,0,1,0,0,1,0,1,0,0,0,1]})

    id  val
0    1    0
1    1    1
2    1    1
3    1    0
4    1    1
5    1    0
6    2    0
7    2    1
8    2    0
9    2    1
10   3    0
11   3    0
12   3    0
13   3    1

As you can see the ID's are duplicated and I have values alternating between 0 and 1. I'd like to somehow calculate the number of times when values switch from 0 to 1. For example like this:

id  val
1    2
2    2
3    1

It's not a count of 1, but rather counter number of times val switch for 0-->1 for each ID.

Answer 1

By using diff

df.groupby('id').val.apply(lambda x : sum(x.diff().eq(1)))
Out[306]: 
id
1    2
2    2
3    1
Name: val, dtype: int64

Answer 2

You could do something like:

>>> # To be more exact: (df['val'] == 1) & (df['val'].shift() == 0
>>> df['val'].diff().eq(1).groupby(df['id']).sum().astype(int)

id
1    2
2    2
3    1

Generally, I like to avoid groupby.apply(...) if at all possible.