How to loop through pandas dataframe and modify value under condition?

Question

I have this pandas dataframe:

df = pd.DataFrame(
    {
    "col1": [1,1,2,3,3,3,4,5,5,5,5]
    }
)
df

I want to add another column that says "last" if the value in col1 doesnt equal the value of col1 in the next row. This is how it should look like:

So far, I can create a column that contains True when if the value in col1 doesnt equal the value of col1 in the next row; and False otherwise:

df["last_row"] = df["col1"].shift(-1)
df['last'] = df["col1"] != df["last_row"]
df = df.drop(["last_row"], axis=1)
df

Now something like

df["last_row"] = df["col1"].shift(-1)
df['last'] = "last" if df["col1"] != df["last_row"]
df = df.drop(["last_row"], axis=1)
df

would be nice, but this is apparently the wrong syntax. How can I manage to do this?

---

Ultimatly, I also want to add numbers that indicate how many time a value appear before this while the last value is always marked with "last". It should look like this:

I'm not sure if this is another step in my development or if this requires a new approach. I read that if I want to loop through an array while modifying values, I should use apply(). However, I don't know how to include conditions in this. Can you help me?

Thanks a lot!

Answer 1

Here's one way. You can obtain a cumulative count based on whether or not the next value in col1 is the same as that of the current row, defining a custom grouper, and taking the DataFrameGroupBy.cumsum. Then add last using a similar criteria using df.shift:

g = df.col1.ne(df.col1.shift(1)).cumsum()
df['update'] = df.groupby(g).cumcount()
ix = df[df.col1.ne(df.col1.shift(-1))].index
# Int64Index([1, 2, 5, 6, 10], dtype='int64')
df.loc[ix,'update'] = 'last'

 col1 update
0      1      0
1      1   last
2      2   last
3      3      0
4      3      1
5      3   last
6      4   last
7      5      0
8      5      1
9      5      2
10     5   last