lockr | A minimal API wrapper for localStorage | Build Tool library

I have the below input

ID LOTYP PROID LOCKR XXXXX 06 01 Y XXXXX 06 02 X XXXXX 06 02 S XXXXX 06 01 R XXXXX 02 01 Y XXXXX 02 02 X XXXXX 02 02 S XXXXX 02 01 R YYYYY 06 01 Y YYYYY 06 02 X YYYYY 06 02 S YYYYY 06 01 R YYYYY 02 01 Y YYYYY 02 02 X YYYYY 02 02 S YYYYY 02 01 R

I would like the below output

ID 0201 0202 0601 0602 XXXXX R, Y S, X R, Y S, X YYYYY R, Y S, X R, Y S, X

I can pivot and listagg separately but I'm struggling to combine them to produce the desired output.

...

I have the below table:

ID LOTYP PROID LOCKR XXXXX 06 01 Y XXXXX 06 02 X XXXXX 06 02 S XXXXX 06 01 R XXXXX 02 01 Y XXXXX 02 02 X XXXXX 02 02 S XXXXX 02 01 R XXXXX 02 01 R XXXXX 02 01 R XXXXX 02 01 R YYYYY 06 01 Y YYYYY 06 02 X YYYYY 06 02 S YYYYY 06 01 R YYYYY 02 01 Y YYYYY 02 02 X YYYYY 02 02 S YYYYY 02 01 R

If I run the below query, it gives the below

...

The problem I'm having is that this set of code NEEDS to happen in order. It is all packaged inside a component that is updating every second.

Here is what I want to happen when the component mounts:

Step 1: On load, retrieve the last known timestamp stored in local storage and subtract it by new Date() /1000 then console log the answer.

Step 2: retrieve all data from local storage and update state (this includes the timestamp), continuing this process every second.

As it stands in my code, step 2 is happening first.

Here's a video of the app I'm working with and the component that's updating every second to provide context to my issue. I highlight my console log being 0. This is the issue I want to fix. I need the console log to not give me 0, but the current timestamp - the previous timestamp. This is so if a user using my app goes offline and comes back, it counts the time they were gone.: https://www.youtube.com/watch?v=N0tOZhHfio4

Here's my current code:

...

If I execute this

...

I'm currently investigating a deadlock issue on a heavily used database but still wasn't able to reproduce it even using concurrent ostress sessions or WHILE 1=1 EXEC StoredProcs. I'd be really glad anyone could shed some light on ways to reproduce it and also help me understand it's behavior.

There are two tables involved and they're partitioned by a hash in a computed column based on the original PK. They were partitioned due to page latch issues on INSERTs in the past.

Thanks to this nice article (Found it to be easier to do this way when doing several times - don't have to issue SELECTs and DBCC PAGE), while trying to reproduce the scenario, I found out that the INSERT statement acquires the following locks:

1. Obj-IS on tConn
2. Obj-IX on tVarConn
3. Page-IX on page from tVarConn
4. Key-RI_NL to the next %%lockres%% from tVarConn
5. Key-X on the %%lockres%% for the row being INSERTed
6. Page-IS on a page from tConn
7. Key-S on a %%lockres%% for the parent key value from tConn

And the DELETE's:

1. Obj-IX on tVarConn
2. Page-IX on tVarConn
3. Key-X on tVarConn (Several of them with different %%lockres%% since I have multiple rows for the nConn/HashID pair)

Plans: Insert Plan Delete Plan

Deadlock graph:

...

I have the following query which returns any locks on a specific table. However, I need it to give more information. I need to know which row is currently being locked.

...