Make Iceberg Dynamic Filter test more resilient

[alexjo2144]

Description

Fixes: #10932

General information

Is this change a fix, improvement, new feature, refactoring, or other?

Flaky test fix

Is this a change to the core query engine, a connector, client library, or the SPI interfaces? (be specific)

Test only

How would you describe this change to a non-technical end user or system administrator?

n/a

Related issues, pull requests, and links

Documentation

(x) No documentation is needed.
( ) Sufficient documentation is included in this PR.
( ) Documentation PR is available with #prnumber.
( ) Documentation issue #issuenumber is filed, and can be handled later.

Release notes

(x) No release notes entries required.
( ) Release notes entries required with the following suggested text:

[alexjo2144]

I'll remove the invocationCount before merge, just wanted to test it a few times on CI

[findepi]

(reminder)

[findepi]

what is [4]?

[losipiuk]

I guess those are stats which correspond to totalprice

[alexjo2144]

Yes, 4 is the fieldId for totalPrice. I looked through the $ metadata tables to see if there was a way to query for that value but couldn't find a good way of doing it

[findepi]

// should find such a price

[losipiuk]

But why are we sure gonna find it?
If we have two files with non overlaping ranges we get empty result range.

E.g:

• f1: [10, 20]
• f2: [3,7]

We get range from the first query: [10, 7]. Which yields empty result from second query.

I am probably missing sth.

[alexjo2144]

That is possible, and would break the test, but seems unlikely. The overlapping range when I ran the test was most of the range of the column. Unless the data is explicitly ordered I think there should be some overlap.

If this still feels too flaky we can take it out and try testing the DF a different way.

[findepi]

why removed?

[alexjo2144]

It is guaranteed by the query now, and driver count stats are kinda flaky.

[aczajkowski]

I had same issue when adding MinIO test's for Iceberg. (this assertion was removed)
Would be nice to merge this PR earlier: #10894 (comment)
than https://github.com/trinodb/trino/pull/10894/checks

[raunaqmorarka]

lgtm % @findepi comments

[alexjo2144]

I removed the invocationCount used for testing, and added another code comment about the assumption of there being an overlapping range in all files. @findepi @losipiuk