Yes I think it is necessary that I also have the same data processing requirements to add a new row to the dataframe

Haha.. congrats on the number.

You can use from_rows. If performance is important, I'd recommend to see if you can still create data in a columnar fashion with the builders.

Dank u.
Ben je Nederlands? Ik spreek beetje Nederlands.

Ik ben Australisch

Bit confused as how to cast of convert to rows to supply an argument to from_rows

Jep, Nederlands. 😄

A Row can have any combination of datatypes. In your case its all f64, but Polars cannot know that normally. So you will have to wrap the f64 data in the AnyValue enum to be able to wrap them inti a Row.

Thanks.

Sorry for being a pest. How do I enumerate row by row?

Sorry for being a pest. How do I enumerate row by row?

How do you mean? What do you want to do?

Let's say we have a dataframe.

How would one go about iterating over it by row?

i.e (this is not the way to do it but will suffice in explaining what I'm trying to do)

for (idx, row) in df.iter().unwrap().enumerate() {
    let { close_time, close_price } = row;
}

You can use get_row to access by index. This however quite slow as the data is layed out columnar in memory. Every row access comprises many cache misses.

Often the same goal can be acquired in a columnar fashion. In you're case, I'd first select the close_time and close_price columns and work on them.

Ideally I would also do the arithmetic on columns.