count the unique occurrence of a col based on other col value

Question

I have the follwing dataframe:

how can I count the unique occurrence of each item when the type=A1? in the above example, it should be 4

I was thinking of something like this:

df['Type']=='A1'
list=df['Item'].unique()
occurance=list.str.len()

is there a better and easier way to do that?

Do not overwrite the built-in `list` function. Also, what is uneasy about the way you are currently doing it? — It_is_Chris, Nov 16 '21 at 21:20
Use [Boolean Indexing](https://pandas.pydata.org/docs/user_guide/10min.html#boolean-indexing) and [nunique](https://pandas.pydata.org/docs/reference/api/pandas.Series.nunique.html) `df.loc[df['Type'] == 'A1', 'Item'].nunique()` as recommended in [this answer](https://stackoverflow.com/a/45760042/15497888) by [Scott Boston](https://stackoverflow.com/users/6361531/scott-boston) — Henry Ecker, Nov 16 '21 at 21:30
Or [groupby nunique](https://pandas.pydata.org/docs/reference/api/pandas.core.groupby.SeriesGroupBy.nunique.html) `df.groupby('Type')['Item'].nunique()` like [this answer](https://stackoverflow.com/a/15411596/15497888) by [Dan Allan](https://stackoverflow.com/users/1221924/dan-allan) if you want all unique item counts for all types. — Henry Ecker, Nov 16 '21 at 21:31

score 1 · Accepted Answer · answered Nov 16 '21 at 21:21

1

Use:

df.groupby('Type')['Item'].nunique()

Output:

Type
A1    4
A2    1

Only A1:

df.groupby('Type')['Item'].nunique()['A1']

Output: 4

answered Nov 16 '21 at 21:21

mozway

score 1 · Answer 2 · answered Nov 16 '21 at 21:21

1

len(df.loc[df["Type"]=="A1"]["Item"].unique())

answered Nov 16 '21 at 21:21

intedgar

score 1 · Answer 3 · answered Nov 16 '21 at 21:21

1

You can use this one liner: df[df['Type']=='A1'][['Item']].drop_duplicates()

Then you can convert the output into list if you need it.

answered Nov 16 '21 at 21:21

Yashar Ahmadov

3 Answers3