Here's the data set:
A 1 5
A 6
A 2
A 3 7
I'd like it to look like this:
A 1 5
A 2 6
A 3 7
A
Note: The first column should not change/move with the sorted columns
Basically I have an ID column that is the parent record number, then clusters of columns that need to be sorted so that within the identical parent rows all values within each cluster sort to the top.
The endpoint I'm loading into ignores values after it sees a blank cell within each cluster, so I need to sort all values to the top within each cluster.
Any ideas? I know I can split out metadata for each and then do a hash join, but we're talking hundreds of clusters, so I'm hoping for a configuration solution.
Thank you!
A 1 5
A 6
A 2
A 3 7
I'd like it to look like this:
A 1 5
A 2 6
A 3 7
A
Note: The first column should not change/move with the sorted columns
Basically I have an ID column that is the parent record number, then clusters of columns that need to be sorted so that within the identical parent rows all values within each cluster sort to the top.
The endpoint I'm loading into ignores values after it sees a blank cell within each cluster, so I need to sort all values to the top within each cluster.
Any ideas? I know I can split out metadata for each and then do a hash join, but we're talking hundreds of clusters, so I'm hoping for a configuration solution.
Thank you!
-
Hi, eneric,
Can you please rephrase your question? I am not exactly sure what is your goal right now. This is how I understand it:
The first column (The one with letters "A") has to be unchanged. The other columns have to be sorted so that all numbers in a row will be sorted in ascending order. But it would imply that for any row, data in it have to be sorted independently and not by the whole columns.
Am I right? And is the number of columns set somehow or you do not know it before the graph run start?
I also do not understand how the numbers move between lines in your example. How is it possible that "2" moved up next to "6"? Is it coincidence that numbers are sorted within rows and also within columns? Which rules apply for this moving?
Thank you in advance.
Best regards, -
Sorry for not being clear! I've expanded the example and attached a screenshot to help illustrate.
The basic structure is preparing for a conversion to xml, so you can think of nesting/sorting the parent elements and the related child elements.
Basically the AAAAA rows need to be sorted and eventually look like the BBBBB rows.
The sort is first by column A (cluster the unique EEIDs), then column B (to sort by date and bring the parent row data to the top), then column D (sort all group 1 elements together and sort blank rows to the bottom), then column F (sort all group 2 elements together and sort blank rows to the bottom).
But each new sort must only be within that group and not change the EEID/Parent record sort, or any of the previous group sorts.
Does this clarify the situation at all?
Thanks! -
Hello, eneric,
No problem, I think I understand now. Please, take a look on this example graph, it should meet your requests.
Best regards,
Please sign in to leave a comment.
Comments 4