def test_table_scan_honors_select_without_case_sensitivity(ts_table):
scan1 = ts_table.new_scan().case_sensitive(False).select(["ID"])
# order of refinements shouldn't matter
scan2 = ts_table.new_scan().select(["ID"]).case_sensitive(False)
expected_schema = Schema([NestedField.required(1, "id", IntegerType.get())])
assert scan1.schema.as_struct() == expected_schema.as_struct()
assert scan2.schema.as_struct() == expected_schema.as_struct()
def bind(self,
struct: StructType,
case_sensitive: bool = True) -> BoundReference:
from iceberg.api import Schema
schema = Schema(struct.fields)
field = schema.find_field(
self.name
) if case_sensitive else schema.case_insensitive_find_field(self.name)
ValidationException.check(field is not None,
"Cannot find field '%s' in struct: %s",
(self.name, schema.as_struct()))
return BoundReference(struct, field)
def prune_columns(file_schema: Schema, expected_schema: Schema) -> List[str]:
"""
Given two Iceberg schema's returns a list of column_names for all id's in the
file schema that are projected in the expected schema
Parameters
----------
file_schema : iceberg.api.Schema
An Iceberg schema of the file being read
expected_schema : iceberg.api.Schema
An Iceberg schema of the final projection
Returns
-------
list
The column names in the file that matched ids in the expected schema
"""
return [column.name for column in file_schema.as_struct().fields
if column.id in get_projected_ids(expected_schema)]
def test_table_scan_honors_select(ts_table):
scan = ts_table.new_scan().select(["id"])
expected_schema = Schema([NestedField.required(1, "id", IntegerType.get())])
assert scan.schema.as_struct() == expected_schema.as_struct()