Yes, I think a transform that allowed you to write SQL queries could be very useful. Hopefully we can use SQLite or DuckDB to do the heavy lifting there.
For the last question: other. Maybe i’m not savy enough, but i have a lot of markdown files with tables. It would be great to be able to read those files e.g. A file with two tables. Open the file. Mark lines by 30 to 77 and 200 to 250 as tables, ignore the rest of the lines and work with the table data. Please do not automate selection of lines. Usually its automatically wrong
Having to manually specify which bits of the .md file are tables sounds rather tedious. Also, it could change every time the input changes. And it should be possible to automatically identify tables (if they are correctly formatted as markdown tables). So I would think it is better to do it automatically? You could perhaps specify something like:
Ability to pin favorite transforms to the top of the sidebar list, or
create my own category = favorites or most used, or
instead of a transforms dropdown for each category (All, Merge, Column, Row, Etc.) have transforms in collapsible sections (lose the dropdown but add a toggle to switch between all or sections view)
Reasoning: I find the full list of transforms+advanced to be a little overwhelming; I only need a few of them (but across different categories and from both simple and advanced lists.) Switching between categories is also a bit of a PITA (I keep clicking “Transform” instead of the dropdown below it, mainly because I’m an idiot with a mouse.
When you know the transaction you intend to use, you don‘t need to select them from sidebar, you just can type the first character of the transformation you want to add and the options will be offered.
I agree that the left pane can be a little overwhelming. But I’m not convinced that duplicating buttons or reordering buttons would be be helpful. You would end up with even more buttons or a changed button order.
That is something we have considered and it is somewhere on the wishlist.
I’m not sure the different categories are distinct enough for that to work well.
I get that. We hide the ‘Advanced’ transforms by default to try to avoid terrifying first time users.
Most customers probably use only a small subset of all the transforms. But it is probably a different subset!
Did you know you can select one or more items in the center pane and just start typing the name of the new transform?
The ability to copy a transforms group or to import saved transform modules may cover my “Other” for now, it being:
It seems often that I need to calculate a statistic then use it in a comparison with each row. Currently this requires creating the stat, which appears in the last row, reproducing it in a new column, copying that up (or down) in the column then removing the line you first created.
Given all the steps after finding the statistic in the first place are consistent regardless of the table structure, could they not be magicked as a tail option for the stat?
I am using the If and Fill items in this post by Anonymous, plus a Filter to remove the row which Stats added, not referring to the Filter used in that particular transform.
If this were an option on the tail of Stats then three or four transforms would be reduced to one. The user burden is not huge but I have found it odd that adding a data column in this fashion after calculating a summary statistic is not already a unitary feature.
