Pandas: How to use method df.name.unique() dynamically to find all unique column entries?

Question

This isn't exactly a question about how to find all unique entries in a column of a dataframe, since I know how I'd do that:

import pandas as pd

df = pd.read_csv('test.txt',delim_whitespace=True)

for key in list(df.keys()):
    uni = set(df[key])

What this is really about, is how to do it with pandas' own methods/functions dynamically and this strange syntax that I can't understand why anyone would use it:

In [101]: list(df.keys())
Out[101]: ['id_cliente', 'id_ordine', 'data_ordine', 'id_medium']

With these keys, you can find their unique column values with the following syntax:

In [102]: df.id_cliente.unique()
Out[102]: array(['madinside', 'lisbeth19'], dtype=object)

I can't use this method dynamically like in my iteration above, can I? I can only use it, if I find out the keys first and manually type in the df.NAME.unique() statement, right?

Why is this a thing? Is this method exclusively intended for interactive use from the python console? Is there a native pandas.DataFrame method for determining unique values dynamically?

BENY · Accepted Answer · 2019-09-20 16:06:16Z

1

You can do it dynamically

df.T.apply(pd.Series.unique,1)

answered Sep 20, 2019 at 16:06

BENY

324k22 gold badges176 silver badges250 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

J.Doe Over a year ago

Thanks. But could you maybe tell me, if the syntax is actually weird or if I'm just weird? Look at my other comment if you want to know what I mean?

SCool · Accepted Answer · 2019-09-20 16:07:47Z

1

Does this work for your df?

unique_stuff = [{col: set(df[col].unique())} for col in df.columns]

edit: actually I don't think you even need a set here. I've removed it below:

unique_stuff  = [{col: df[col].unique().tolist()} for col in df.columns]

edited Sep 20, 2019 at 16:07

answered Sep 20, 2019 at 16:02

SCool

3,4366 gold badges31 silver badges68 bronze badges

1 Comment

J.Doe Over a year ago

Yeah I could do that, it's just that I found it weird that you can access a column of a dataframe by using it's column name string as a code fragment. Doesn't this mean, that when a DataFrame instance is created, the column names are used to create static variables, where the name of the static variables is also the name of the column? Am I the only one who finds this weird? stackoverflow.com/questions/68645/…

Collectives™ on Stack Overflow

Pandas: How to use method df.name.unique() dynamically to find all unique column entries?

2 Answers 2

1 Comment

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related