MCPcopy Create free account
hub / github.com/apache/arrow / test_inference

Method test_inference

python/pyarrow/tests/test_csv.py:1697–1716  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

1695 self.check_reader(reader, expected_schema, [])
1696
1697 def test_inference(self):
1698 # Inference is done on first block
1699 rows = b"a,b\n123,456\nabc,de\xff\ngh,ij\n"
1700 expected_schema = pa.schema([('a', pa.string()),
1701 ('b', pa.binary())])
1702
1703 read_options = ReadOptions()
1704 read_options.block_size = len(rows)
1705 reader = self.open_bytes(rows, read_options=read_options)
1706 self.check_reader(reader, expected_schema,
1707 [{'a': ['123', 'abc', 'gh'],
1708 'b': [b'456', b'de\xff', b'ij']}])
1709
1710 read_options.block_size = len(rows) - 1
1711 reader = self.open_bytes(rows, read_options=read_options)
1712 self.check_reader(reader, expected_schema,
1713 [{'a': ['123', 'abc'],
1714 'b': [b'456', b'de\xff']},
1715 {'a': ['gh'],
1716 'b': [b'ij']}])
1717
1718 def test_inference_failure(self):
1719 # Inference on first block, then conversion failure on second block

Callers

nothing calls this directly

Calls 6

open_bytesMethod · 0.95
check_readerMethod · 0.95
ReadOptionsClass · 0.90
lenFunction · 0.85
schemaMethod · 0.45
stringMethod · 0.45

Tested by

no test coverage detected