Perf. improvements with activation and cell merge.

[MayaKirova]

Closes #16146

Split pipes in two, where one does the merging and the other tracks active rows and unmerges within related merge sequences.

Additional information (check all that apply):

• Bug fix
• New functionality
• Documentation
• Demos
• CI/CD

Checklist:

• All relevant tags have been applied to this PR
• This PR includes unit tests covering all the new code (test guidelines)
• This PR includes API docs for newly added methods/properties (api docs guidelines)
• This PR includes feature/README.MD updates for the feature docs
• This PR includes general feature table updates in the root README.MD
• This PR includes CHANGELOG.MD updates for newly added functionality
• This PR contains breaking changes
• This PR includes ng update migrations for the breaking changes (migrations guidelines)
• This PR includes behavioral changes and the feature specification has been updated with them

[dkamburov]

Openned rowPinning sample and pinned a row:

Also when I pin a row from cell merge sample I'm getting the following violation:

[MayaKirova]

@dkamburov First one, I was not able to reproduce:

Are you sure you have the latest changes in the branch or are there any additional steps needed to reproduce?

Second one is a sample issue. Because of the way the data is added, the same object refs are used in the grid sample:

for (let i = 1; i <= 600_000; i++) {
            const rnd = Math.floor(Math.random() * length);
            allData.push(INVOICE_DATA[rnd]);
        }

Hence, duplicate keys.

I fixed it in the sample.

[dkamburov]

@MayaKirova appologize on the non reproducible sample: I do had some changes on that sample(didn't noticed them at first as they were from my previous testing on the cell merging):

[cellMergeMode]="'always'" to the grid and [merge]="true" to the columns

[mddragnev]

Additional, issues I found:

1. If you click on any of the headers in the dev sample an error is thrown.

2. Go to the performance project and enable the merging on all columns. Run the performance samples using npm run start:performance and go to the grid with 1 million records. Sort the Position column. Click on the first cell of the Registered column. An error is thrown with maximum callstack exceeded.

3. For 1 million records on the performance sample, it takes 4 to 5 seconds to merge the cells. Sometimes the browser crashes out of memory. This makes all other features that are pipe related to feel very slow.

[mddragnev]

Could we use a regular for-of loop here and instead of creating 3 new array coppies we can mutate the result (which is cloned beforehand) in place?

[MayaKirova]

Do you mean going through all items in res and replacing each one in the original array with the unmerged version?

[mddragnev]

Yes.

[MayaKirova]

I mean, we could. But it would look a bit ugly and also not sure how much more performant it would be considering that in the particular scenario:

2. Go to the performance project and enable the merging on all columns. Run the performance samples using npm run start:performance and go to the grid with 1 million records. Sort the Position column. Click on the first cell of the Registered column. An error is thrown with maximum callstack exceeded.

It would need to loop through 300k+ records and replace each one.

[mddragnev]

I thought that

for(let i =0;i<res.length;++i) {
    result[index+i] = res[i];
}

would be faster because there is no array copying but changes the result in place. However, seems like the slice and concat version is perhaps more optimized by the js engine. I have tested this and, indeed, your current version is faster.

[dkamburov]

mouse wheel when there is cell merging is becoming choppy, it was faster before the changes and there is a difference when compared to a grid without cell merging
https://github.com/user-attachments/assets/da5b5e05-a0aa-48da-a9ae-6bd464a9584b

[MayaKirova]

mouse wheel when there is cell merging is becoming choppy, it was faster before the changes and there is a difference when compared to a grid without cell merging https://github.com/user-attachments/assets/da5b5e05-a0aa-48da-a9ae-6bd464a9584b

@dkamburov I can confirm that none of the merge related pipes are actually triggered on mouse wheel, so it doesn't make sense for the change to effect it.

Are you sure it's not simply because of the large amount of data in the grid from the sample? It's choppy for me even without merging.