Yes, the duplicate branch is helpful for working within transforms. But to copy a string from one transform to another would be a nice to have item. I use certain transforms as template files, and would like to combine parts and pieces between them when needed.
The issue is: if you copy part of a chain of transforms without the input(s), all the column related options in the transforms copied will be nuked when they are pasted without the input item (because there are no longer any columns). Still thinking about whether there is a good way around that.
We use the current approach because it makes it easy to have more than 1 version installed. Running the uninstaller on the old version should clean up everything. But we will consider a change of approach in future.
Undo/Redo is high on the wishlist, but needs to be carefully thought through so EDT doesn’t end up using all the RAM on your computer.
Multi-threading will only help if you have parallel branches in your .transform (not a single linear sequence). Is that the case? How long is the .transform taking to run now?
I certainly understand there are considerations probably in all the options you wrote above. In most of my work I have paralel branches because I normally use EDT with multiple sources and several branches. Allow me also to ask some more questions about some of the other ideas:
what would come up on the support for multiple screens? I’m not sure I understand that one.
what integration with other tools are you guys thinking about?
I would also say that having more liberty to move nodes and branches around the screen would be a nice to have. Or to expand the working area.
Multiple screens - Some customers have asked to have the option to have the right pane displayed on a second monitor.
Integration - We don’t have any prior agenda on this one. it depends on what existing systems customers are integrating EDT into.
Hopefully that is already easy to do. Just drag a ‘rubber box’ around the items you want to move and drag them. The canvas should expand as needed.
Frequency calculation to count unique items in selected columns (cities, zip codes, sex, whatever) with output in columns such as City | Count of City | zip codes | Count of zip codes | sex | Count of sex | whatever | Count of whatever
Support for web input would be grand
Calculation/transform output displayed in a resizable window allowing selection and copying. The window could be called from the choicebox (in the blue square in the capture) or with a button (mockup in the green square) without the need to open Excel, Numbers or whatever app is associated with CSV
You can use Count to count the number of unique values in a column. Each Count only works on one column though. You could merge the results, but I am not sure it would make sense to merge multiple Count results into 1 dataset.
Do you have a particular web app in mind?
You can copy from the existing table. Just select the row, column or cells and Ctrl+C (Windows) or Cmd-C (Mac). Admittedly, it isn’t particularly obvious. Or did you have something else in mind?
How about the ability to create templates? For example, I usually start my transformations with the following transforms, Sample, Header, Whitespace, and summary.
I know we can use the awesome batching function, but templates can be useful.