How do I remove all non numerical values in a column

[shan312]

Hello, I am newbie and just started on the platform help is much appreciated thank you.

[Kathrin]

Hi @shan312 and welcome to the KNIME Community Forum!

You can use the Column Filter node to remove all non numerical columns, by using the Type Selection and including only numerical data types:

Cheers
Kathrin

[shan312]

Hello sorry how do we transfer the columns from excluded to included thanks.

[Kathrin]

Hi,

to include / exclude columns by data type you first need to select Type Selection in the upper part. Afterwards you can use the checkboxes to select the data types you want to include. The columns are then automatically moved to the include frame.

Cheers
Kathrin

PS: In case some of your numerical columns are at the moment of type string you can convert the date type with the String to Number node.

[shan312]

Hi, after selecting the data type I want to include how do I include only certain columns on my side the columns are all in the excluded side. Thanks.

[Kathrin]

Hi,

with the column filter node you can either filter columns based on the data type or manually, but you can’t do both at the same time.

In case some of your numerical columns are not included after selecting all numerical data types (integer, double, and long) probably the data type of some of your columns is not set correctly. The current data type of a column is shown via the little icon next to the column header. To change the data type before using the column filter node you can use for example the string to number node to convert the data type.

Cheers
Kathrin

[Thyme]

Hi @shan312,
if you want to remove non-numerical values within a column, you can either filter the rows or replace them with missing values, both via a RegEx condition. Is that what you want to do?
This RegEx might be too basic for you, as it only allows unsigned integers.

exclude non-numerical values.knwf (16.1 KB)

[Kathrin]

Ahh sorry, I misunderstood the question. Thank you @Thyme for providing a correct suggestion!