Make UPDATE respect non nullability of columns

[homar]

Description

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

a fix

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

core engine

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

it makes update queries respect non nullable columns and not insert nulls into them

Related issues, pull requests, and links

Fixes: #13435

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

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

# Core
* Makes Update queries respect non nullability of columns. ({https://github.com/trinodb/trino/issues/13435}`13435`)

[homar]

failure is the same as here #13658

[homar]

TestDeltaLakeDatabricksInsertCompatibility.testCompressionWithOptimizedWriter seems to be a transient issue

And singlestore failed because of #13677 (review) which should be fixed already

[losipiuk]

declare as Map

[losipiuk]

i -> i -> index -> index. Or static import on Function.identity()

[losipiuk]

Please keep old variable name: updatedColumnValuesBuilder

[losipiuk]

It feels the iteration would be simpler if you just iterate in parallel over orderedColumValues and targetColumnNames. IIUC those have same length so you can just do the iteration using integer index from 0..len(orderedColumValues).
The you would just look at columns you are interested in and look up for extra information (about nullability) by looking into tableMetadata. Currently you are iterating over all the columns and skipping once you are not interested in. This is counterintutive and requires using updatedColumnNameToIndex.

[homar]

So actually i thought about that but how would you look into tableMetadata for specific column ?
There is getColumn method but its implementation doesn't seem very efficient:

public ColumnMetadata getColumn(String name)
    {
        return getColumns().stream()
                .filter(columnMetadata -> columnMetadata.getName().equals(name))
                .findFirst()
                .orElseThrow(() -> new IllegalArgumentException("Invalid column name: " + name));
    }

We may end up going over that column list many many times. WDYT @losipiuk

[losipiuk]

Can you just cache it as a name->CM map?

[homar]

@martint can we merge it ? or do you still have some concerns ?