def test_query_avro(con, test_data_dir, tmp_db):
hdfs_path = pjoin(test_data_dir, 'avro/tpch_region_avro')
avro_schema = {
"fields": [
{"type": ["int", "null"], "name": "R_REGIONKEY"},
{"type": ["string", "null"], "name": "R_NAME"},
{"type": ["string", "null"], "name": "R_COMMENT"},
],
"type": "record",
"name": "a",
}
table = con.avro_file(hdfs_path, avro_schema, database=tmp_db)
qualified_name = table.op().name
_, name, _ = fully_qualified_re.match(qualified_name).groups()
# table exists
assert name in con.list_tables(database=tmp_db)
expr = table.r_name.value_counts()
expr.execute()
assert table.count().execute() == 5
df = table.execute()
assert len(df) == 5
def _match_name(self):
m = fully_qualified_re.match(self._qualified_name)
if not m:
raise com.IbisError(
'Cannot determine database name from {}'.format(
self._qualified_name))
db, quoted, unquoted = m.groups()
return db, quoted or unquoted
def list_tables(self, like=None, database=None):
statement = 'SHOW TABLES'
if database is not None:
statement += f' IN {database}'
if like:
m = fully_qualified_re.match(like)
if m:
database, quoted, unquoted = m.groups()
like = quoted or unquoted
return self.list_tables(like=like, database=database)
statement += f" LIKE '{like}'"
return self._filter_with_like(
[row[0] for row in self.raw_sql(statement).fetchall()])
def rename(self, new_name, database=None):
"""Rename table inside Impala.
References to the old table are no longer valid.
"""
m = fully_qualified_re.match(new_name)
if not m and database is None:
database = self._database
statement = RenameTable(
self._qualified_name, new_name, new_database=database
)
self._client.raw_sql(statement)
op = self.op().change_name(statement.new_qualified_name)
return type(self)(op)
def _match_name(self):
m = fully_qualified_re.match(self._qualified_name)
if not m:
return None, self._qualified_name
db, quoted, unquoted = m.groups()
return db, quoted or unquoted