Inset value in polars dataframe

Python Help
1

Hi

I have a polars dataframe where I need to update values in a column where the rows corresponds to a column called ID (type str). Kind of like this quasi code:
if row fulfills ID = ‘string’ update a float value in column ‘columnname’

I’m fairly new to polars dataframe and can’t get it to work.

2

I think you need something in the direction of

df=df.with_column(pl.col('name').where(id==Val).then(new).else(old))

3

Thanks

Your reply is confusing to me… Isn’t with_column for adding a column to the dataframe? Isn’t there some way of updating a single row column, i.e. a single cell. I couldn’t get you code working.

4

No, you can use with_column to overwrite a column, if you specify a column that has the same name as a column you already have.

You need a polars.when().then().otherwise() structure. Which was the very first thing on top of my google results when I searched for “polars where then else”.

https://docs.pola.rs/api/python/stable/reference/expressions/api/polars.when.html

It genuinly is hard to get polars to work when you’re starting out. The expressions are tricky. Sometimes AI chatbots help.

5

Okay thanks!

Now I’m able to update values. However, when I try to repeat and update more values after the first update previously updated values are changed to the original values.

name = 'Column2'
String1 = 'STR1'
new_df = new_df.with_columns(pl.when(pl.col('ID') == String1).then(1).alias(name))


String2 = 'STR2'

new_df = new_df.with_columns(pl.when(pl.col('ID') == String2).then(2).alias(name))
print(new_df)

EDIT: I have another question linked to this. I use this to add a column, which should be a float, but is a i32. Could that be the reason why i fail…?

df.with_columns(name = pl.lit(None))
6

I think my approach was wrong. I was able to do what I wanted to do by first creating a dict with key and value and form a dataframe from the dict. Then I could join the two dataframes of key.