Openrefine: Transform All Columns

[allthebits]

When Transforming ALL rows of data:

add a checkbox in the popup that allows you to "select all columns" to apply the transformation to all rows of data and all columns of data regardless of the column name, preventing the column names from being statically defined within the JSON representation of the transformation.

This request is based on the following discussion thread from the OpenRefine google group

[jenyoung]

I'd also like to see the list of the Common Transforms replicated here. Thanks!

[thadguidry]

Done ! This was released as part of OpenRefine 2.7 ! Use the latest OpenRefine version. Closing this issue now.

[ostephens]

I think this request is subtly different to what has been implemented.

The current implementation does the following:

• Allows you to write a transformation
• Allow you to select the columns it applies to
• Applies that transformation to each column selected, using the column name as the key

This results in multiple items in the project history with one entry per column in the project. When applying this to other projects, the transformations will only be applied if the columns have exactly the same name as in the original project.

This request, as I understand it, is to allow the user to select to apply the transformation to 'All columns' which would iterate through all columns in the project and apply the transformation to each column regardless of name. This would create a single item in the History which was to 'apply transform to all columns' and would iterate through all columns in a project regardless of their name.

This would enable the transformation to be applied to future projects regardless of the column names

[thadguidry]

@ostephens I see. Then we probably want some good feature name for this, because we will need a menu item name for it. Hmm.

[ostephens]

I think it would be an additional option on this screen:

So the existing "Transform" option under the 'All' menu would still be used, you'd just have another selection option on this screen "Apply to all columns regardless of name" (or something)

[thadguidry]

@wetneb How's our Spark prototype architecture for handling this use case that Owen mentions? No concerns, do-able?

[wetneb]

No concerns, this is independent from the change of architecture.