Fix distinct hash table resizing

[acquamarin]

Closes #3339
This PR solves a distinct hash table issue:
When we append values to distinct hash table, we forgot to resize the hashtable even if it is full. This causes an infinite loop issue.
This PR solves the issue simply by resizing the distinct hashtable upon each insertion.

[ted-wq-x]

This change can indeed solve the infinite loop issue,but the current hash slots has a very poor performance.

[acquamarin]

This change can indeed solve the infinite loop issue,but the current hash slots has a very poor performance.

Since the query has distinct keyword in agg keys (count(distinct a)), we can only run the query in single thread mode. That's the reason why you were experiencing a slow spped.

[ted-wq-x]

I understand distinct only run in single thread mode, but the query in another graph db only need 2.8s (also single thread mode), in kuzu the execution didn‘t finished in 5min .

[ray6080]

I understand distinct only run in single thread mode, but the query in another graph db only need 2.8s (also single thread mode), in kuzu the execution didn‘t finished in 5min .

Looks like we might have a performance bug regarding to this then. Will look into it. Thanks!

[acquamarin]

#3339

I opened a new issue about performance on distinct hash aggregation: #3349