如何在pydatatable中输入数据框列？

Question

I'm trying to explore datatypes of a frame in pydatatable.

Here I have a dataframe:

ventas_duda_dt = dt.Frame(
    {"cust_id":[893232.34],"sales":['$123,4532.93'],"profit_perc":['10%']})

and its types:

ventas_duda_dt.stypes

and the datatypes of each of these fields as:

(stype.float64, stype.str32, stype.str32)

Now I would like to cast three of these columns to int, float, float from float, str, str , here. How it can be achieved by removing these special symbols ($,,,%) in field values?

I was going though the documentation, but I couldn't find a reference over there respect to this topic.

Answer 1

In datatable the easiest way to change a column's type is to assign the new type to the column:

ventas_duda_dt["cust_id"] = int
ventas_duda_dt["sales"] = float
ventas_duda_dt["profit_perc"] = float

However, in your example the situation is more complicated because the string values are not directly convertible. You'd need special converters that know about percent signs, comma-separated numbers, currency symbols, etc. Such functionality doesn't exist in datatable yet, but it may be possible to add it in the future, provided that you can specify what exactly such converters should do.