Fix numNodes in CopyNode

[ray6080]

There was a change (https://github.com/kuzudb/kuzu/pull/2070/files#r1348216726) to fix edge case for copying empty csv file, which doesn't fully fix the problem with numNodes.

[codecov]

Codecov Report

All modified lines are covered by tests ✅

Comparison is base (4aa1d07) 89.65% compared to head (c6116f1) 89.62%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2161      +/-   ##
==========================================
- Coverage   89.65%   89.62%   -0.04%     
==========================================
  Files         989      989              
  Lines       35740    35740              
==========================================
- Hits        32042    32031      -11     
- Misses       3698     3709      +11     

Files	Coverage Δ
src/processor/operator/persistent/copy_node.cpp	95.91% <100.00%> (ø)

... and 3 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Riolku]

Curious, what problems does the current code give rise to?

[ray6080]

Curious, what problems does the current code give rise to?

The case is that when sharedState->sharedNodeGroup is null, we would set numNodes to 0, which can be incorrect. sharedState->sharedNodeGroup is used only when threads have tuples left to write out. So it could be that each thread writes out all their tuples, while we still set numNodes to 0. Does this make sense?

[Riolku]

Uh, is NodeGroup a fixed number of nodes stored together on disk? And so if all threads happen to write out a number of nodes that is a multiple of that number we get a totally incorrect statistic?

[ray6080]

Uh, is NodeGroup a fixed number of nodes stored together on disk? And so if all threads happen to write out a number of nodes that is a multiple of that number we get a totally incorrect statistic?

Yep.