def test_to_bytes(self):
self.assertEqual(b'\x00\x00', Literal.of(False).to_byte_buffer())
self.assertEqual(b'\x01\x00', Literal.of(True).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00',
Literal.of(1234).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00\x00\x00\x00\x00',
Literal.of(1234).to(LongType.get()).to_byte_buffer())
self.assertEqual(b'\x19\x04\x9e?', Literal.of(1.2345).to_byte_buffer())
self.assertEqual(
b'\x8d\x97\x6e\x12\x83\xc0\xf3\x3f',
Literal.of(1.2345).to(DoubleType.get()).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00',
Literal.of(1234).to(DateType.get()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(TimeType.get()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(
TimestampType.with_timezone()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(
TimestampType.without_timezone()).to_byte_buffer())
self.assertEqual(b'foo', Literal.of("foo").to_byte_buffer())
self.assertEqual(
b'\xf7\x9c>\tg|K\xbd\xa4y?4\x9c\xb7\x85\xe7',
Literal.of(uuid.UUID(
"f79c3e09-677c-4bbd-a479-3f349cb785e7")).to_byte_buffer())
self.assertEqual(b'foo', Literal.of(bytes(b'foo')).to_byte_buffer())
self.assertEqual(b'foo',
Literal.of(bytearray(b'foo')).to_byte_buffer())
def add_timestamp_expression(self, timestamp, expr_func):
if isinstance(timestamp, str):
timestamp = Literal.of(timestamp).to(
TimestampType.without_timezone()).value / 1000
self.add_timestamp_filter(expr_func("timestamp_ms", timestamp))
return self
def test_timestamp_without_zone_human_string():
ts_tz = TimestampType.without_timezone()
identity = Transforms.identity(ts_tz)
ts_str = "2017-12-01T10:12:55.038194"
ts = Literal.of(ts_str).to(ts_tz)
assert identity.to_human_string(ts.value) == ts_str
def test_from_bytes(self):
self.assertEqual(
False, Conversions.from_byte_buffer(BooleanType.get(), b'\x00'))
self.assertEqual(
True, Conversions.from_byte_buffer(BooleanType.get(), b'\x01'))
self.assertEqual(
1234,
Conversions.from_byte_buffer(IntegerType.get(),
b'\xd2\x04\x00\x00'))
self.assertEqual(
1234,
Conversions.from_byte_buffer(LongType.get(),
b'\xd2\x04\x00\x00\x00\x00\x00\x00'))
self.assertAlmostEqual(1.2345,
Conversions.from_byte_buffer(
FloatType.get(), b'\x19\x04\x9e?'),
places=5)
self.assertAlmostEqual(
1.2345,
Conversions.from_byte_buffer(DoubleType.get(),
b'\x8d\x97\x6e\x12\x83\xc0\xf3\x3f'))
self.assertEqual(
1234,
Conversions.from_byte_buffer(DateType.get(), b'\xd2\x04\x00\x00'))
self.assertEqual(
100000000000,
Conversions.from_byte_buffer(TimeType.get(),
b'\x00\xe8vH\x17\x00\x00\x00'))
self.assertEqual(
100000000000,
Conversions.from_byte_buffer(TimestampType.with_timezone(),
b'\x00\xe8vH\x17\x00\x00\x00'))
self.assertEqual(
100000000000,
Conversions.from_byte_buffer(TimestampType.without_timezone(),
b'\x00\xe8vH\x17\x00\x00\x00'))
self.assertEqual(
"foo", Conversions.from_byte_buffer(StringType.get(), b'foo'))
self.assertEqual(
uuid.UUID("f79c3e09-677c-4bbd-a479-3f349cb785e7"),
Conversions.from_byte_buffer(
UUIDType.get(), b'\xf7\x9c>\tg|K\xbd\xa4y?4\x9c\xb7\x85\xe7'))
self.assertEqual(
b'foo', Conversions.from_byte_buffer(FixedType.of_length(3),
b'foo'))
self.assertEqual(
b'foo', Conversions.from_byte_buffer(BinaryType.get(), b'foo'))
self.assertEqual(
Decimal(123.45).quantize(Decimal(".01")),
Conversions.from_byte_buffer(DecimalType.of(5, 2), b'\x30\x39'))
self.assertEqual(
Decimal(123.4567).quantize(Decimal(".0001")),
Conversions.from_byte_buffer(DecimalType.of(5, 4),
b'\x00\x12\xd6\x87'))
self.assertEqual(
Decimal(-123.4567).quantize(Decimal(".0001")),
Conversions.from_byte_buffer(DecimalType.of(5, 4),
b'\xff\xed\x29\x79'))
def test_partition_spec(self):
schema = Schema(NestedField.required(1, "i", IntegerType.get()),
NestedField.required(2, "l", LongType.get()),
NestedField.required(3, "d", DateType.get()),
NestedField.required(4, "t", TimeType.get()),
NestedField.required(5, "ts", TimestampType.without_timezone()),
NestedField.required(6, "dec", DecimalType.of(9, 2)),
NestedField.required(7, "s", StringType.get()),
NestedField.required(8, "u", UUIDType.get()),
NestedField.required(9, "f", FixedType.of_length(3)),
NestedField.required(10, "b", BinaryType.get()))
specs = [PartitionSpec.builder_for(schema).identity("i").build(),
PartitionSpec.builder_for(schema).identity("l").build(),
PartitionSpec.builder_for(schema).identity("d").build(),
PartitionSpec.builder_for(schema).identity("t").build(),
PartitionSpec.builder_for(schema).identity("ts").build(),
PartitionSpec.builder_for(schema).identity("dec").build(),
PartitionSpec.builder_for(schema).identity("s").build(),
PartitionSpec.builder_for(schema).identity("u").build(),
PartitionSpec.builder_for(schema).identity("f").build(),
PartitionSpec.builder_for(schema).identity("b").build(),
PartitionSpec.builder_for(schema).bucket("i", 128).build(),
PartitionSpec.builder_for(schema).bucket("l", 128).build(),
PartitionSpec.builder_for(schema).bucket("d", 128).build(),
PartitionSpec.builder_for(schema).bucket("t", 128).build(),
PartitionSpec.builder_for(schema).bucket("ts", 128).build(),
PartitionSpec.builder_for(schema).bucket("dec", 128).build(),
PartitionSpec.builder_for(schema).bucket("s", 128).build(),
PartitionSpec.builder_for(schema).bucket("u", 128).build(),
PartitionSpec.builder_for(schema).bucket("f", 128).build(),
PartitionSpec.builder_for(schema).bucket("b", 128).build(),
PartitionSpec.builder_for(schema).year("d").build(),
PartitionSpec.builder_for(schema).month("d").build(),
PartitionSpec.builder_for(schema).day("d").build(),
PartitionSpec.builder_for(schema).year("ts").build(),
PartitionSpec.builder_for(schema).month("ts").build(),
PartitionSpec.builder_for(schema).day("ts").build(),
PartitionSpec.builder_for(schema).hour("ts").build(),
PartitionSpec.builder_for(schema).truncate("i", 10).build(),
PartitionSpec.builder_for(schema).truncate("l", 10).build(),
PartitionSpec.builder_for(schema).truncate("dec", 10).build(),
PartitionSpec.builder_for(schema).truncate("s", 10).build(),
PartitionSpec.builder_for(schema).add_without_field_id(6, "dec_unsupported", "unsupported").build(),
PartitionSpec.builder_for(schema).add(6, 1111, "dec_unsupported", "unsupported").build(),
]
for spec in specs:
self.assertEqual(spec, TestHelpers.round_trip_serialize(spec))
def test_bucket_hash(self):
buckets = [
[Transforms.bucket(IntegerType.get(), 100), 34, 2017239379],
[Transforms.bucket(LongType.get(), 100), 34, 2017239379],
[Transforms.bucket(DateType.get(), 100), 17486, -653330422],
[Transforms.bucket(TimeType.get(), 100), 81068000000, -662762989],
[Transforms.bucket(TimestampType.without_timezone(), 100), 1510871468000000, -2047944441],
[Transforms.bucket(DecimalType.of(9, 2), 100), decimal.Decimal("14.20"), -500754589],
[Transforms.bucket(StringType.get(), 100), "iceberg", 1210000089],
[Transforms.bucket(UUIDType.get(), 100), uuid.UUID("f79c3e09-677c-4bbd-a479-3f349cb785e7"), 1488055340],
[Transforms.bucket(FixedType.of_length(3), 128), b'foo', -156908512],
[Transforms.bucket(BinaryType.get(), 128), b'\x00\x01\x02\x03', -188683207]
]
for bucket in buckets:
self.assertEqual(bucket[2], bucket[0].hash(bucket[1]))
def rg_expected_schema():
return Schema([
NestedField.required(1, "string_col", StringType.get()),
NestedField.required(2, "long_col", LongType.get()),
NestedField.required(3, "int_col", IntegerType.get()),
NestedField.optional(4, "float_col", FloatType.get()),
NestedField.optional(5, "null_col", StringType.get()),
NestedField.optional(6, "missing_col", StringType.get()),
NestedField.optional(7, "no_stats_col", StringType.get()),
NestedField.optional(8, "ts_wtz_col", TimestampType.with_timezone()),
NestedField.optional(9, "ts_wotz_col",
TimestampType.without_timezone()),
NestedField.optional(10, "big_decimal_type", DecimalType.of(38, 5)),
NestedField.optional(11, "small_decimal_type", DecimalType.of(10, 2)),
NestedField.optional(12, "date_type", DateType.get()),
])
def test_string_to_timestamp_literal():
timestamp_str = Literal.of("2017-08-18T14:21:01.919+00:00")
timestamp = timestamp_str.to(TimestampType.with_timezone())
avro_val = avro_conversion["long-timestamp-micros"](
dateutil.parser.parse("2017-08-18T14:21:01.919+00:00"), None)
assert avro_val == timestamp.value
timestamp_str = Literal.of("2017-08-18T14:21:01.919")
timestamp = timestamp_str.to(TimestampType.without_timezone())
assert avro_val == timestamp.value
timestamp_str = Literal.of("2017-08-18T14:21:01.919-07:00")
timestamp = timestamp_str.to(TimestampType.with_timezone())
avro_val = avro_conversion["long-timestamp-micros"](
dateutil.parser.parse("2017-08-18T21:21:01.919+00:00"), None)
assert avro_val == timestamp.value
def test_primitive_types(primitive_type_test_parquet_file):
expected_schema = Schema([
NestedField.required(1, "int_col", IntegerType.get()),
NestedField.optional(2, "bigint_col", LongType.get()),
NestedField.optional(3, "str_col", StringType.get()),
NestedField.optional(4, "float_col", FloatType.get()),
NestedField.optional(5, "dbl_col", DoubleType.get()),
NestedField.optional(6, "decimal_col", DecimalType.of(9, 2)),
NestedField.optional(7, "big_decimal_col", DecimalType.of(19, 5)),
NestedField.optional(8, "huge_decimal_col", DecimalType.of(38, 9)),
NestedField.optional(9, "date_col", DateType.get()),
NestedField.optional(10, "ts_col", TimestampType.without_timezone()),
NestedField.optional(11, "ts_wtz_col", TimestampType.with_timezone()),
NestedField.optional(12, "bool_col", BooleanType.get())
])
compare_schema(
expected_schema,
convert_parquet_to_iceberg(primitive_type_test_parquet_file))
def remove_time_filters(self, expressions, expression):
if expression.op == Operation.AND:
self.remove_time_filters(expressions, expression.left)
self.remove_time_filters(expressions, expression.right)
return
elif isinstance(expression, UnboundPredicate):
pred = expression
ref = pred.ref
lit = pred.lit
if ref.name in ScanSummaryBuilder.TIMESTAMP_NAMES:
ts_literal = lit.to(TimestampType.without_timezone())
millis = ScanSummaryBuilder.to_millis(ts_literal.value)
self.add_timestamp_filter(
Expressions.predicate(pred.op, "timestamp_ms", millis))
return
expressions.append(expression)
def test_to_bytes(self):
self.assertEqual(b'\x00', Literal.of(False).to_byte_buffer())
self.assertEqual(b'\x01', Literal.of(True).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00',
Literal.of(1234).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00\x00\x00\x00\x00',
Literal.of(1234).to(LongType.get()).to_byte_buffer())
self.assertEqual(b'\x19\x04\x9e?', Literal.of(1.2345).to_byte_buffer())
self.assertEqual(
b'\x8d\x97\x6e\x12\x83\xc0\xf3\x3f',
Literal.of(1.2345).to(DoubleType.get()).to_byte_buffer())
self.assertEqual(b'\xd2\x04\x00\x00',
Literal.of(1234).to(DateType.get()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(TimeType.get()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(
TimestampType.with_timezone()).to_byte_buffer())
self.assertEqual(
b'\x00\xe8vH\x17\x00\x00\x00',
Literal.of(100000000000).to(
TimestampType.without_timezone()).to_byte_buffer())
self.assertEqual(b'foo', Literal.of("foo").to_byte_buffer())
self.assertEqual(
b'\xf7\x9c>\tg|K\xbd\xa4y?4\x9c\xb7\x85\xe7',
Literal.of(uuid.UUID(
"f79c3e09-677c-4bbd-a479-3f349cb785e7")).to_byte_buffer())
self.assertEqual(b'foo', Literal.of(bytes(b'foo')).to_byte_buffer())
self.assertEqual(b'foo',
Literal.of(bytearray(b'foo')).to_byte_buffer())
# Decimal on 2-bytes
self.assertEqual(
b'\x30\x39',
Literal.of(123.45).to(DecimalType.of(5, 2)).to_byte_buffer())
# Decimal on 3-bytes to test that we use the minimum number of bytes
self.assertEqual(
b'\x12\xd6\x87',
Literal.of(123.4567).to(DecimalType.of(7, 4)).to_byte_buffer())
# Negative decimal to test two's complement
self.assertEqual(
b'\xed\x29\x79',
Literal.of(-123.4567).to(DecimalType.of(7, 4)).to_byte_buffer())
def test_basic_read(primitive_type_test_file, pyarrow_primitive_array,
pyarrow_schema):
expected_schema = Schema([
NestedField.required(1, "int_col", IntegerType.get()),
NestedField.optional(2, "bigint_col", LongType.get()),
NestedField.optional(3, "str_col", StringType.get()),
NestedField.optional(4, "float_col", FloatType.get()),
NestedField.optional(5, "dbl_col", DoubleType.get()),
NestedField.optional(6, "decimal_col", DecimalType.of(9, 2)),
NestedField.optional(7, "big_decimal_col", DecimalType.of(19, 5)),
NestedField.optional(8, "huge_decimal_col", DecimalType.of(38, 9)),
NestedField.optional(9, "date_col", DateType.get()),
NestedField.optional(10, "ts_col", TimestampType.without_timezone()),
NestedField.optional(11, "ts_wtz_col", TimestampType.with_timezone()),
NestedField.optional(12, "bool_col", BooleanType.get())
])
input_file = FileSystemInputFile(get_fs(primitive_type_test_file, conf={}),
primitive_type_test_file, {})
reader = ParquetReader(input_file, expected_schema, {},
Expressions.always_true(), True)
source_table = pa.table(pyarrow_primitive_array, schema=pyarrow_schema)
assert reader.read() == source_table
def test_timestamp_without_zone_with_zone_in_literal():
with raises(RuntimeError):
timestamp_str = Literal.of("2017-08-18T14:21:01.919+07:00")
timestamp_str.to(TimestampType.without_timezone())
def test_to_json_conversion():
spec_schema = Schema(NestedField.required(1, "i", IntegerType.get()),
NestedField.required(2, "l", LongType.get()),
NestedField.required(3, "d", DateType.get()),
NestedField.required(4, "t", TimeType.get()),
NestedField.required(5, "ts", TimestampType.without_timezone()),
NestedField.required(6, "dec", DecimalType.of(9, 2)),
NestedField.required(7, "s", StringType.get()),
NestedField.required(8, "u", UUIDType.get()),
NestedField.required(9, "f", FixedType.of_length(3)),
NestedField.required(10, "b", BinaryType.get()))
specs = [
PartitionSpec.builder_for(spec_schema).identity("i").build(),
PartitionSpec.builder_for(spec_schema).identity("l").build(),
PartitionSpec.builder_for(spec_schema).identity("d").build(),
PartitionSpec.builder_for(spec_schema).identity("t").build(),
PartitionSpec.builder_for(spec_schema).identity("ts").build(),
PartitionSpec.builder_for(spec_schema).identity("dec").build(),
PartitionSpec.builder_for(spec_schema).identity("s").build(),
PartitionSpec.builder_for(spec_schema).identity("u").build(),
PartitionSpec.builder_for(spec_schema).identity("f").build(),
PartitionSpec.builder_for(spec_schema).identity("b").build(),
PartitionSpec.builder_for(spec_schema).bucket("i", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("l", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("d", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("t", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("ts", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("dec", 128).build(),
PartitionSpec.builder_for(spec_schema).bucket("s", 128).build(),
PartitionSpec.builder_for(spec_schema).year("d").build(),
PartitionSpec.builder_for(spec_schema).month("d").build(),
PartitionSpec.builder_for(spec_schema).day("d").build(),
PartitionSpec.builder_for(spec_schema).year("ts").build(),
PartitionSpec.builder_for(spec_schema).month("ts").build(),
PartitionSpec.builder_for(spec_schema).day("ts").build(),
PartitionSpec.builder_for(spec_schema).hour("ts").build(),
PartitionSpec.builder_for(spec_schema).truncate("i", 10).build(),
PartitionSpec.builder_for(spec_schema).truncate("l", 10).build(),
PartitionSpec.builder_for(spec_schema).truncate("dec", 10).build(),
PartitionSpec.builder_for(spec_schema).truncate("s", 10).build(),
PartitionSpec.builder_for(spec_schema).add(6, "dec_bucket", "bucket[16]").build()
]
expected_spec_strs = [
"[\n i: identity(1)\n]",
"[\n l: identity(2)\n]",
"[\n d: identity(3)\n]",
"[\n t: identity(4)\n]",
"[\n ts: identity(5)\n]",
"[\n dec: identity(6)\n]",
"[\n s: identity(7)\n]",
"[\n u: identity(8)\n]",
"[\n f: identity(9)\n]",
"[\n b: identity(10)\n]",
"[\n i_bucket: bucket[128](1)\n]",
"[\n l_bucket: bucket[128](2)\n]",
"[\n d_bucket: bucket[128](3)\n]",
"[\n t_bucket: bucket[128](4)\n]",
"[\n ts_bucket: bucket[128](5)\n]",
"[\n dec_bucket: bucket[128](6)\n]",
"[\n s_bucket: bucket[128](7)\n]",
"[\n d_year: year(3)\n]",
"[\n d_month: month(3)\n]",
"[\n d_day: day(3)\n]",
"[\n ts_year: year(5)\n]",
"[\n ts_month: month(5)\n]",
"[\n ts_day: day(5)\n]",
"[\n ts_hour: hour(5)\n]",
"[\n i_truncate: truncate[10](1)\n]",
"[\n l_truncate: truncate[10](2)\n]",
"[\n dec_truncate: truncate[10](6)\n]",
"[\n s_truncate: truncate[10](7)\n]",
"[\n dec_bucket: bucket[16](6)\n]",
]
for (spec, expected_spec_str) in zip(specs, expected_spec_strs):
assert str(spec) == expected_spec_str
def test_byte_buffer_conversions(self):
# booleans are stored as 0x00 for 'false' and a non-zero byte for 'true'
self.assertConversion(False, BooleanType.get(), b'\x00')
self.assertConversion(True, BooleanType.get(), b'\x01')
self.assertEqual(b'\x00', Literal.of(False).to_byte_buffer())
self.assertEqual(b'\x01', Literal.of(True).to_byte_buffer())
# integers are stored as 4 bytes in little-endian order
# 84202 is 0...01|01001000|11101010 in binary
# 11101010 -> 234 (-22), 01001000 -> 72, 00000001 -> 1, 00000000 -> 0
self.assertConversion(84202, IntegerType.get(), bytes([234, 72, 1, 0]))
self.assertEqual(bytes([234, 72, 1, 0]),
Literal.of(84202).to_byte_buffer())
# longs are stored as 8 bytes in little-endian order
# 200L is 0...0|11001000 in binary
# 11001000 -> 200 (-56), 00000000 -> 0, ... , 00000000 -> 0
self.assertConversion(200, LongType.get(),
bytes([200, 0, 0, 0, 0, 0, 0, 0]))
self.assertEqual(bytes([200, 0, 0, 0, 0, 0, 0, 0]),
Literal.of(200).to(LongType.get()).to_byte_buffer())
# floats are stored as 4 bytes in little-endian order
# floating point numbers are represented as sign * 2ˆexponent * mantissa
# -4.5F is -1 * 2ˆ2 * 1.125 and encoded as 11000000|10010000|0...0 in binary
# 00000000 -> 0, 00000000 -> 0, 10010000 -> 144 (-112), 11000000 -> 192 (-64),
self.assertConversion(-4.5, FloatType.get(), bytes([0, 0, 144, 192]))
self.assertEqual(bytes([0, 0, 144, 192]),
Literal.of(-4.5).to_byte_buffer())
# doubles are stored as 8 bytes in little-endian order
# floating point numbers are represented as sign * 2ˆexponent * mantissa
# 6.0 is 1 * 2ˆ4 * 1.5 and encoded as 01000000|00011000|0...0
# 00000000 -> 0, ... , 00011000 -> 24, 01000000 -> 64
self.assertConversion(6.0, DoubleType.get(),
bytes([0, 0, 0, 0, 0, 0, 24, 64]))
self.assertEqual(bytes([0, 0, 0, 0, 0, 0, 24, 64]),
Literal.of(6.0).to(DoubleType.get()).to_byte_buffer())
# dates are stored as days from 1970-01-01 in a 4-byte little-endian int
# 1000 is 0...0|00000011|11101000 in binary
# 11101000 -> 232 (-24), 00000011 -> 3, ... , 00000000 -> 0
self.assertConversion(1000, DateType.get(), bytes([232, 3, 0, 0]))
self.assertEqual(bytes([232, 3, 0, 0]),
Literal.of(1000).to(DateType.get()).to_byte_buffer())
# time is stored as microseconds from midnight in an 8-byte little-endian long
# 10000L is 0...0|00100111|00010000 in binary
# 00010000 -> 16, 00100111 -> 39, ... , 00000000 -> 0
self.assertConversion(10000, TimeType.get(),
bytes([16, 39, 0, 0, 0, 0, 0, 0]))
self.assertEqual(
bytes([16, 39, 0, 0, 0, 0, 0, 0]),
Literal.of(10000).to(LongType.get()).to(
TimeType.get()).to_byte_buffer())
# timestamps are stored as microseconds from 1970-01-01 00:00:00.000000 in an 8-byte little-endian long
# 400000L is 0...110|00011010|10000000 in binary
# 10000000 -> 128 (-128), 00011010 -> 26, 00000110 -> 6, ... , 00000000 -> 0
self.assertConversion(400000, TimestampType.without_timezone(),
bytes([128, 26, 6, 0, 0, 0, 0, 0]))
self.assertConversion(400000, TimestampType.with_timezone(),
bytes([128, 26, 6, 0, 0, 0, 0, 0]))
self.assertEqual(
bytes([128, 26, 6, 0, 0, 0, 0, 0]),
Literal.of(400000).to(LongType.get()).to(
TimestampType.without_timezone()).to_byte_buffer())
self.assertEqual(
bytes([128, 26, 6, 0, 0, 0, 0, 0]),
Literal.of(400000).to(LongType.get()).to(
TimestampType.with_timezone()).to_byte_buffer())
# strings are stored as UTF-8 bytes (without length)
# 'A' -> 65, 'B' -> 66, 'C' -> 67
self.assertConversion("ABC", StringType.get(), bytes([65, 66, 67]))
self.assertEqual(bytes([65, 66, 67]),
Literal.of("ABC").to_byte_buffer())
# uuids are stored as 16-byte big-endian values
# f79c3e09-677c-4bbd-a479-3f349cb785e7 is encoded as F7 9C 3E 09 67 7C 4B BD A4 79 3F 34 9C B7 85 E7
# 0xF7 -> 11110111 -> 247 (-9), 0x9C -> 10011100 -> 156 (-100), 0x3E -> 00111110 -> 62,
# 0x09 -> 00001001 -> 9, 0x67 -> 01100111 -> 103, 0x7C -> 01111100 -> 124,
# 0x4B -> 01001011 -> 75, 0xBD -> 10111101 -> 189 (-67), 0xA4 -> 10100100 -> 164 (-92),
# 0x79 -> 01111001 -> 121, 0x3F -> 00111111 -> 63, 0x34 -> 00110100 -> 52,
# 0x9C -> 10011100 -> 156 (-100), 0xB7 -> 10110111 -> 183 (-73), 0x85 -> 10000101 -> 133 (-123),
# 0xE7 -> 11100111 -> 231 (-25)
self.assertConversion(
uuid.UUID("f79c3e09-677c-4bbd-a479-3f349cb785e7"), UUIDType.get(),
bytes([
247, 156, 62, 9, 103, 124, 75, 189, 164, 121, 63, 52, 156, 183,
133, 231
]))
self.assertEqual(
bytes([
247, 156, 62, 9, 103, 124, 75, 189, 164, 121, 63, 52, 156, 183,
133, 231
]),
Literal.of(uuid.UUID(
"f79c3e09-677c-4bbd-a479-3f349cb785e7")).to_byte_buffer())
# fixed values are stored directly
# 'a' -> 97, 'b' -> 98
self.assertConversion(bytes("ab", "utf8"), FixedType.of_length(2),
bytes([97, 98]))
self.assertEqual(bytes([97, 98]),
Literal.of(bytes("ab", "utf8")).to_byte_buffer())
# binary values are stored directly
# 'Z' -> 90
self.assertConversion(bytearray("Z", "utf8"), BinaryType.get(),
bytes([90]))
self.assertEqual(bytes([90]),
Literal.of(bytearray("Z", "utf8")).to_byte_buffer())
# decimals are stored as unscaled values in the form of two's-complement big-endian binary,
# using the minimum number of bytes for the values
# 345 is 0...1|01011001 in binary
# 00000001 -> 1, 01011001 -> 89
self.assertConversion(
Decimal(3.45).quantize(Decimal(".01")), DecimalType.of(3, 2),
bytes([1, 89]))
self.assertEqual(
bytes([1, 89]),
Literal.of(3.45).to(DecimalType.of(3, 2)).to_byte_buffer())
# decimal on 3-bytes to test that we use the minimum number of bytes and not a power of 2
# 1234567 is 00010010|11010110|10000111 in binary
# 00010010 -> 18, 11010110 -> 214, 10000111 -> 135
self.assertConversion(
Decimal(123.4567).quantize(Decimal(".0001")), DecimalType.of(7, 4),
bytes([18, 214, 135]))
self.assertEqual(
bytes([18, 214, 135]),
Literal.of(123.4567).to(DecimalType.of(7, 4)).to_byte_buffer())
# negative decimal to test two's complement
# -1234567 is 11101101|00101001|01111001 in binary
# 11101101 -> 237, 00101001 -> 41, 01111001 -> 121
self.assertConversion(
Decimal(-123.4567).quantize(Decimal(".0001")),
DecimalType.of(7, 4), bytes([237, 41, 121]))
self.assertEqual(
bytes([237, 41, 121]),
Literal.of(-123.4567).to(DecimalType.of(7, 4)).to_byte_buffer())
# test empty byte in decimal
# 11 is 00001011 in binary
# 00001011 -> 11
self.assertConversion(
Decimal(0.011).quantize(Decimal(".001")), DecimalType.of(10, 3),
bytes([11]))
self.assertEqual(
bytes([11]),
Literal.of(0.011).to(DecimalType.of(10, 3)).to_byte_buffer())
@pytest.mark.parametrize("test_input,test_type,expected",
[(1, BucketFloat(100), -142385009),
(1, BucketDouble(100), -142385009)])
def test_spec_values_dbl(test_input, test_type, expected):
assert test_type.hash(test_input) == expected
@pytest.mark.parametrize(
"test_input,test_type,scale_factor,expected",
[(Decimal("14.20"), DecimalType.of(9, 2), Decimal(10)**-2, -500754589),
(Decimal("137302769811943318102518958871258.37580"), DecimalType.of(
38, 5), Decimal(10)**-5, -32334285)])
def test_spec_values_dec(test_input, test_type, scale_factor, expected):
getcontext().prec = 38
assert Bucket.get(test_type,
100).hash(test_input.quantize(scale_factor)) == expected
@pytest.mark.parametrize("test_input,test_type,expected", [
(Literal.of("2017-11-16").to(DateType.get()), DateType.get(), -653330422),
(Literal.of("22:31:08").to(TimeType.get()), TimeType.get(), -662762989),
(Literal.of("2017-11-16T22:31:08").to(TimestampType.without_timezone()),
TimestampType.without_timezone(), -2047944441),
(Literal.of("2017-11-16T14:31:08-08:00").to(TimestampType.with_timezone()),
TimestampType.with_timezone(), -2047944441)
])
def test_spec_values_datetime_uuid(test_input, test_type, expected):
assert Bucket.get(test_type, 100).hash(test_input.value) == expected
class AvroToIceberg(object):
FIELD_ID_PROP = "field-id"
FIELD_TYPE_PROP = "type"
FIELD_NAME_PROP = "name"
FIELD_LOGICAL_TYPE_PROP = "logicalType"
FIELD_FIELDS_PROP = "fields"
FIELD_ITEMS_PROP = "items"
FIELD_ELEMENT_ID_PROP = "element-id"
AVRO_JSON_PRIMITIVE_TYPES = ("boolean", "int", "long", "float", "double", "bytes", "string")
AVRO_JSON_COMPLEX_TYPES = ("record", "array", "enum", "fixed")
TYPE_PROCESSING_MAP = {str: lambda x, y: AvroToIceberg.convert_str_type(x, y),
dict: lambda x, y: AvroToIceberg.convert_complex_type(x, y),
list: lambda x, y: AvroToIceberg.convert_union_type(x, y)}
COMPLEX_TYPE_PROCESSING_MAP = {"record": lambda x, y: AvroToIceberg.convert_record_type(x, y),
"array": lambda x, y: AvroToIceberg.convert_array_type(x, y),
"map": lambda x, y: AvroToIceberg.convert_map_type(x, y)}
PRIMITIVE_FIELD_TYPE_MAP = {"boolean": BooleanType.get(),
"bytes": BinaryType.get(),
"date": DateType.get(),
"double": DoubleType.get(),
"float": FloatType.get(),
"int": IntegerType.get(),
"long": LongType.get(),
"string": StringType.get(),
"time-millis": TimeType.get(),
"timestamp-millis": TimestampType.without_timezone()}
PROCESS_FUNCS = {TypeID.STRUCT: lambda avro_row, field: AvroToIceberg.get_field_from_struct(avro_row, field),
TypeID.LIST: lambda avro_row, field: AvroToIceberg.get_field_from_list(avro_row, field),
TypeID.MAP: lambda avro_row, field: AvroToIceberg.get_field_from_map(avro_row, field)}
@staticmethod
def convert_avro_schema_to_iceberg(avro_schema):
if avro_schema.get(AvroToIceberg.FIELD_TYPE_PROP) != "record":
raise RuntimeError("Cannot convert avro schema to iceberg %s" % avro_schema)
struct = AvroToIceberg.convert_type(avro_schema, None)
return Schema(struct[0].fields)
@staticmethod
def convert_record_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
if avro_field_type != "record":
raise RuntimeError("Field type muse be 'record': %s" % avro_field_type)
fields = avro_field.get(AvroToIceberg.FIELD_FIELDS_PROP)
iceberg_fields = []
if next_id is None:
next_id = len(fields)
for field in fields:
iceberg_field, next_id = AvroToIceberg.convert_avro_field_to_iceberg(field, next_id=next_id)
iceberg_fields.append(iceberg_field)
return StructType.of(iceberg_fields), next_id
@staticmethod
def convert_avro_field_to_iceberg(field, next_id):
field_type, is_optional, next_id = AvroToIceberg.convert_type(field, next_id)
if field.get(AvroToIceberg.FIELD_ID_PROP) is None:
return field_type, next_id
if is_optional:
return NestedField.optional(field.get(AvroToIceberg.FIELD_ID_PROP),
field.get(AvroToIceberg.FIELD_NAME_PROP),
field_type), next_id
else:
return NestedField.required(field.get(AvroToIceberg.FIELD_ID_PROP),
field.get(AvroToIceberg.FIELD_NAME_PROP),
field_type), next_id
@staticmethod
def convert_type(field, next_id=None):
avro_field_type = field.get(AvroToIceberg.FIELD_TYPE_PROP)
optional = AvroToIceberg.is_option_schema(avro_field_type)
processing_func = AvroToIceberg.TYPE_PROCESSING_MAP.get(type(avro_field_type))
if processing_func is None:
raise RuntimeError("No function found to process %s" % avro_field_type)
iceberg_type, next_id = processing_func(field, next_id)
return iceberg_type, optional, next_id
@staticmethod
def convert_str_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
logical_type = avro_field.get(AvroToIceberg.FIELD_LOGICAL_TYPE_PROP)
if not isinstance(avro_field_type, str):
raise RuntimeError("Field type must be of type str: %s" % avro_field_type)
if avro_field_type in AvroToIceberg.AVRO_JSON_PRIMITIVE_TYPES:
if logical_type is not None:
return AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP.get(logical_type), next_id
else:
return AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP.get(avro_field_type), next_id
elif avro_field_type in AvroToIceberg.AVRO_JSON_COMPLEX_TYPES:
if logical_type is not None:
processing_func = AvroToIceberg.COMPLEX_TYPE_PROCESSING_MAP.get(logical_type)
else:
processing_func = AvroToIceberg.COMPLEX_TYPE_PROCESSING_MAP.get(avro_field_type)
if processing_func is None:
raise RuntimeError("No function found to process %s" % avro_field_type)
return processing_func(avro_field, next_id)
else:
raise RuntimeError("Unknown type %s" % avro_field_type)
@staticmethod
def convert_complex_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
if not isinstance(avro_field_type, dict):
raise RuntimeError("Complex field type must be of type dict: %s" % avro_field_type)
return AvroToIceberg.convert_avro_field_to_iceberg(avro_field_type, next_id)
@staticmethod
def convert_union_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
if not isinstance(avro_field_type, list):
raise RuntimeError("Union field type must be of type list: %s" % avro_field_type)
if len(avro_field_type) > 2:
raise RuntimeError("Cannot process unions larger than 2 items: %s" % avro_field_type)
for item in avro_field_type:
if isinstance(item, str) and item == "null":
continue
avro_field_type = item
avro_field[AvroToIceberg.FIELD_TYPE_PROP] = avro_field_type
items = AvroToIceberg.convert_type(avro_field, next_id)
return items[0], items[2]
@staticmethod
def convert_array_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
if avro_field_type != "array":
raise RuntimeError("Avro type must be array: %s" % avro_field_type)
element_id = avro_field.get(AvroToIceberg.FIELD_ELEMENT_ID_PROP)
items = avro_field.get(AvroToIceberg.FIELD_ITEMS_PROP)
is_optional = AvroToIceberg.is_option_schema(items)
if isinstance(items, str) and items in AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP:
item_type = AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP.get(items)
if item_type is None:
raise RuntimeError("No mapping found for type %s" % items)
else:
raise RuntimeError("Complex list types not yet implemented")
if is_optional:
return ListType.of_optional(element_id, item_type), next_id
else:
return ListType.of_required(element_id, item_type), next_id
@staticmethod
def convert_map_type(avro_field, next_id=None):
avro_field_type = avro_field.get(AvroToIceberg.FIELD_TYPE_PROP)
avro_logical_type = avro_field.get(AvroToIceberg.FIELD_LOGICAL_TYPE_PROP)
if avro_field_type != "array" or avro_logical_type != "map":
raise RuntimeError("Avro type must be array and logical type must be map: %s" % avro_logical_type)
is_optional = False
items = avro_field.get(AvroToIceberg.FIELD_ITEMS_PROP)
for field in items.get(AvroToIceberg.FIELD_FIELDS_PROP, list()):
if field.get(AvroToIceberg.FIELD_NAME_PROP) == "key":
key_id = field.get(AvroToIceberg.FIELD_ID_PROP)
if not isinstance(field.get(AvroToIceberg.FIELD_TYPE_PROP), str):
raise RuntimeError("Support for complex map keys not yet implemented")
key_type = AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP.get(field.get(AvroToIceberg.FIELD_TYPE_PROP))
elif field.get(AvroToIceberg.FIELD_NAME_PROP) == "value":
value_id = field.get(AvroToIceberg.FIELD_ID_PROP)
if not isinstance(field.get(AvroToIceberg.FIELD_TYPE_PROP), str):
raise RuntimeError("Support for complex map values not yet imeplemented")
value_type = AvroToIceberg.PRIMITIVE_FIELD_TYPE_MAP.get(field.get(AvroToIceberg.FIELD_TYPE_PROP))
if is_optional:
return MapType.of_optional(key_id, value_id, key_type, value_type), next_id
else:
return MapType.of_required(key_id, value_id, key_type, value_type), next_id
@staticmethod
def is_option_schema(field_type):
if isinstance(field_type, list) and len(field_type) == 2 and "null" in field_type:
return True
return False
@staticmethod
def read_avro_file(iceberg_schema, data_file):
fo = data_file.new_fo()
avro_reader = fastavro.reader(fo)
for avro_row in avro_reader:
iceberg_row = dict()
for field in iceberg_schema.as_struct().fields:
iceberg_row[field.name] = AvroToIceberg.get_field_from_avro(avro_row, field)
yield iceberg_row
fo.close()
@staticmethod
def read_avro_row(iceberg_schema, avro_reader):
try:
for avro_row in avro_reader:
iceberg_row = dict()
for field in iceberg_schema.as_struct().fields:
iceberg_row[field.name] = AvroToIceberg.get_field_from_avro(avro_row, field)
yield iceberg_row
except StopIteration:
return
@staticmethod
def get_field_from_avro(avro_row, field):
try:
return AvroToIceberg.PROCESS_FUNCS.get(field.type.type_id,
AvroToIceberg.get_field_from_primitive)(avro_row, field)
except KeyError:
raise RuntimeError("Don't know how to get field of type: %s" % field.type.type_id)
@staticmethod
def get_field_from_primitive(avro_row, field):
try:
return avro_row[field.name]
except KeyError:
if field.is_required:
raise RuntimeError("Field is required but missing in source %s\n%s:" % (field, avro_row))
@staticmethod
def get_field_from_struct(avro_row, field):
field_obj = {}
for nested_field in field.type.fields:
field_obj[nested_field.name] = AvroToIceberg.get_field_from_avro(avro_row[field.name], nested_field)
return field_obj
@staticmethod
def get_field_from_list(avro_row, field):
try:
return avro_row[field.name]
except KeyError:
if field.is_required:
raise RuntimeError("Field is required but missing in source %s\n%s:" % (field, avro_row))
@staticmethod
def get_field_from_map(avro_row, field):
val_map = dict()
try:
avro_value = avro_row[field.name]
except KeyError:
if field.is_required:
raise RuntimeError("Field is required but missing in source %s\n%s:" % (field, avro_row))
else:
return None
for val in avro_value:
val_map[val['key']] = val['value']
return val_map
import pyarrow as pa
from pyarrow.parquet import lib, ParquetFile
_logger = logging.getLogger(__name__)
arrow_type_map = {lib.Type_BOOL: lambda x=None: BooleanType.get(),
lib.Type_DATE32: lambda x=None: DateType.get(),
lib.Type_DECIMAL128: lambda x=None: DecimalType.of(x.precision, x.scale),
lib.Type_DOUBLE: lambda x=None: DoubleType.get(),
lib.Type_FIXED_SIZE_BINARY: lambda x=None: FixedType.of_length(x.byte_width),
lib.Type_BINARY: lambda x=None: BinaryType.get(),
lib.Type_FLOAT: lambda x=None: FloatType.get(),
lib.Type_STRING: lambda x=None: StringType.get(),
lib.Type_INT32: lambda x=None: IntegerType.get(),
lib.Type_INT64: lambda x=None: LongType.get(),
lib.Type_TIMESTAMP: lambda x=None: (TimestampType.without_timezone()
if x.tz is None
else TimestampType.with_timezone())
}
def get_nested_field(field_id: int, field_name: str, field_type: Type, nullable: bool) -> NestedField:
if nullable:
return NestedField.optional(field_id, field_name, field_type)
else:
return NestedField.required(field_id, field_name, field_type)
def get_list_field(col_type: pa.ListType) -> ListType:
if col_type.value_field.nullable:
return ListType.of_optional(int(col_type.value_field.metadata[b"PARQUET:field_id"].decode("utf-8")),
])
def not_eq_rewrite(request):
yield request.param
@pytest.fixture(scope="session",
params=[
Literal.of(False),
Literal.of(34),
Literal.of(35),
Literal.of(36.75),
Literal.of(8.75),
Literal.of("2017-11-29").to(DateType.get()),
Literal.of("11:30:0").to(TimeType.get()),
Literal.of("2017-11-29T11:30:07.123").to(
TimestampType.without_timezone()),
Literal.of("2017-11-29T11:30:07.123+01:00").to(
TimestampType.with_timezone()),
Literal.of("abc"),
Literal.of(uuid.uuid4()),
Literal.of(bytes([0x01, 0x02,
0x03])).to(FixedType.of_length(3)),
Literal.of(bytes([0x03, 0x04, 0x05,
0x06])).to(BinaryType.get()),
Literal.of(Decimal(122.50).quantize(Decimal(".01")))
])
def literal(request):
yield request.param
@pytest.fixture(scope="session",