def test_set_index_twice(sample_column_names, sample_inferred_logical_types):
schema = TableSchema(sample_column_names, sample_inferred_logical_types)
schema.set_index(None)
assert schema.index is None
schema.set_time_index(None)
assert schema.time_index is None
schema = TableSchema(
sample_column_names,
sample_inferred_logical_types,
index="id",
time_index="signup_date",
)
original_schema = schema.get_subset_schema(list(schema.columns.keys()))
schema.set_index("id")
assert schema.index == "id"
assert schema.semantic_tags["id"] == {"index"}
assert schema == original_schema
schema.set_time_index("signup_date")
assert schema.time_index == "signup_date"
assert schema.semantic_tags["signup_date"] == {"time_index"}
assert schema == original_schema
def test_get_subset_schema_all_params(sample_column_names,
sample_inferred_logical_types):
# The first element is self, so it won't be included in kwargs
possible_schema_params = inspect.getfullargspec(
TableSchema.__init__)[0][1:]
kwargs = {
"column_names": sample_column_names,
"logical_types": {
**sample_inferred_logical_types, "email": EmailAddress
},
"name": "test_dt",
"index": "id",
"time_index": "signup_date",
"semantic_tags": {
"age": "test_tag"
},
"table_metadata": {
"created_by": "user1"
},
"column_metadata": {
"phone_number": {
"format": "xxx-xxx-xxxx"
}
},
"use_standard_tags": False,
"column_descriptions": {
"age": "this is a description"
},
"column_origins": "base",
"validate": True,
}
# Confirm all possible params to TableSchema init are present with non-default values where possible
assert set(possible_schema_params) == set(kwargs.keys())
schema = TableSchema(**kwargs)
copy_schema = schema.get_subset_schema(sample_column_names)
assert schema == copy_schema
assert schema is not copy_schema
def test_schema_rename(sample_column_names, sample_inferred_logical_types):
table_metadata = {"table_info": "this is text"}
id_description = "the id of the row"
id_origin = "base"
schema = TableSchema(
sample_column_names,
sample_inferred_logical_types,
index="id",
time_index="signup_date",
table_metadata=table_metadata,
column_descriptions={"id": id_description},
column_origins={"id": id_origin},
)
original_schema = schema.get_subset_schema(list(schema.columns.keys()))
renamed_schema = schema.rename({"age": "birthday"})
# Confirm original schema hasn't changed
assert schema == original_schema
assert "age" not in renamed_schema.columns
assert "birthday" in renamed_schema.columns
# confirm that metadata and descriptions are there
assert renamed_schema.metadata == table_metadata
assert schema.columns["id"].description == id_description
assert schema.columns["id"].origin == id_origin
old_col = schema.columns["age"]
new_col = renamed_schema.columns["birthday"]
assert old_col.logical_type == new_col.logical_type
assert old_col.semantic_tags == new_col.semantic_tags
swapped_schema = schema.rename({"age": "full_name", "full_name": "age"})
swapped_back_schema = swapped_schema.rename({
"age": "full_name",
"full_name": "age"
})
assert swapped_back_schema == schema
def test_get_subset_schema(sample_column_names, sample_inferred_logical_types):
schema = TableSchema(sample_column_names, sample_inferred_logical_types)
new_schema = schema.get_subset_schema(sample_column_names[1:4])
for col in new_schema.columns:
assert new_schema.semantic_tags[col] == schema.semantic_tags[col]
assert new_schema.logical_types[col] == schema.logical_types[col]
