Here the challenge is that how to replace the file paths in the recursion tree.
This must be tested, so in theory, we would have a plan with multiple joins.
For particular join, there are the left and the right.
Left and right have ReadRel and we should have a way to update that plan.
So what would happen is, we will have a Substrait plan, then we go through the recursion tree, then for each table
we lookup the iceberg_catalog and load the said table and extract the files via the pyiceberg.table.Table.scan().
Here the challenge is that how to replace the file paths in the recursion tree. This must be tested, so in theory, we would have a plan with multiple joins. For particular join, there are the left and the right.
Left and right have
ReadRel
and we should have a way to update that plan.For instance we could do the following;
So what would happen is, we will have a Substrait plan, then we go through the recursion tree, then for each table we lookup the
iceberg_catalog
and load the said table and extract the files via thepyiceberg.table.Table.scan()
.