impala/testdata/datasets/functional/schema_constraints.csv at dedff5e25c19da724ae032a52354cd4fd61ec40a

mirror of https://github.com/apache/impala.git synced 2026-01-21 15:03:35 -05:00

Files

Vuk Ercegovac db98dc6504 IMPALA-4993: extend dictionary filtering to collections

Currently, top-level scalar columns in parquet files can
be used at runtime to prune row-groups by evaluating certain
conjuncts over the column's dictionary (if available).

This change extends such pruning to scalar values that are
stored in collection type columns. Currently, dictionary
pruning works by finding eligible conjuncts for top-level
slots. Since only top-level slots are supported, the slots
are implicitly part of the scan node's tuple descriptor.
With this change, we track eligible conjuncts by slot as well
as the tuple that contains the slot (either top-level or
nested collection). Since collection conjuncts are already
managed by a map that associates tuple descriptors to a list
of their conjuncts, this extension follows the existing
representation.

The frontend builds the mapping of SlotId to conjuncts that
are dictionary filterable. This mapping now includes SlotId's
that reference nested tuples. The backend is adjusted to
use the same representation. In addition, collection
readers are decomposed into scalar filterable columns and
other, non-dictionary filterable readers. When filtering
a row group using a conjunct associated to a (possibly)
nested collection type, an additional tuple buffer is
allocated per tuple descriptor.

Testing:
- e2e test extended to illustrate row-groups that are pruned
  by nested collection dictionary filters.

Change-Id: If3a2abcfc3d0f7d18756816659fed77ce12668dd
Reviewed-on: http://gerrit.cloudera.org:8080/8775
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Impala Public Jenkins

2018-01-19 20:37:25 +00:00

13 KiB

Raw Blame History

1	# Table level constraints:
2	# Allows for defining constraints on which file formats to generate for an individual
3	# table. The table name should match the base table name defined in the schema template
4	# file.
5	table_name:stringids, constraint:restrict_to, table_format:hbase/none/none
6	table_name:hbasecolumnfamilies, constraint:restrict_to, table_format:hbase/none/none
7	table_name:insertalltypesagg, constraint:restrict_to, table_format:hbase/none/none
8	table_name:alltypessmallbinary, constraint:restrict_to, table_format:hbase/none/none
9	table_name:insertalltypesaggbinary, constraint:restrict_to, table_format:hbase/none/none
10	table_name:hbasealltypeserror, constraint:restrict_to, table_format:hbase/none/none
11	table_name:hbasealltypeserrornonulls, constraint:restrict_to, table_format:hbase/none/none
12	table_name:alltypesinsert, constraint:restrict_to, table_format:text/none/none
13	table_name:stringpartitionkey, constraint:restrict_to, table_format:text/none/none
14	table_name:alltypesnopart_insert, constraint:restrict_to, table_format:text/none/none
15	table_name:insert_overwrite_nopart, constraint:restrict_to, table_format:text/none/none
16	table_name:insert_overwrite_partitioned, constraint:restrict_to, table_format:text/none/none
17	table_name:insert_string_partitioned, constraint:restrict_to, table_format:text/none/none
18	table_name:alltypesinsert, constraint:restrict_to, table_format:parquet/none/none
19	table_name:alltypesnopart_insert, constraint:restrict_to, table_format:parquet/none/none
20	table_name:alltypesinsert, constraint:restrict_to, table_format:text/none/none
21	table_name:alltypesnopart_insert, constraint:restrict_to, table_format:text/none/none
22	table_name:insert_overwrite_nopart, constraint:restrict_to, table_format:text/none/none
23	table_name:insert_overwrite_partitioned, constraint:restrict_to, table_format:text/none/none
24	table_name:insert_string_partitioned, constraint:restrict_to, table_format:text/none/none
25	table_name:alltypesinsert, constraint:restrict_to, table_format:parquet/none/none
26	table_name:alltypesnopart_insert, constraint:restrict_to, table_format:parquet/none/none
27	table_name:insert_overwrite_nopart, constraint:restrict_to, table_format:parquet/none/none
28	table_name:insert_overwrite_partitioned, constraint:restrict_to, table_format:parquet/none/none
29	table_name:insert_string_partitioned, constraint:restrict_to, table_format:parquet/none/none
30	table_name:old_rcfile_table, constraint:restrict_to, table_format:rc/none/none
31	table_name:bad_text_lzo, constraint:restrict_to, table_format:text/lzo/block
32	table_name:bad_text_gzip, constraint:restrict_to, table_format:text/gzip/block
33	table_name:bad_seq_snap, constraint:restrict_to, table_format:seq/snap/block
34	table_name:bad_avro_snap_strings, constraint:restrict_to, table_format:avro/snap/block
35	table_name:bad_avro_snap_floats, constraint:restrict_to, table_format:avro/snap/block
36	table_name:bad_avro_decimal_schema, constraint:restrict_to, table_format:avro/snap/block
37	table_name:bad_parquet, constraint:restrict_to, table_format:parquet/none/none
38	table_name:bad_parquet_strings_negative_len, constraint:restrict_to, table_format:parquet/none/none
39	table_name:bad_parquet_strings_out_of_bounds, constraint:restrict_to, table_format:parquet/none/none
40	table_name:bad_magic_number, constraint:restrict_to, table_format:parquet/none/none
41	table_name:bad_metadata_len, constraint:restrict_to, table_format:parquet/none/none
42	table_name:bad_dict_page_offset, constraint:restrict_to, table_format:parquet/none/none
43	table_name:bad_compressed_size, constraint:restrict_to, table_format:parquet/none/none
44	table_name:alltypesagg_hive_13_1, constraint:restrict_to, table_format:parquet/none/none
45	table_name:kite_required_fields, constraint:restrict_to, table_format:parquet/none/none
46	table_name:bad_column_metadata, constraint:restrict_to, table_format:parquet/none/none
47	table_name:lineitem_multiblock, constraint:restrict_to, table_format:parquet/none/none
48	table_name:lineitem_sixblocks, constraint:restrict_to, table_format:parquet/none/none
49	table_name:lineitem_multiblock_one_row_group, constraint:restrict_to, table_format:parquet/none/none
50	table_name:customer_multiblock, constraint:restrict_to, table_format:parquet/none/none
51	# TODO: Support Avro. Data loading currently fails for Avro because complex types
52	# cannot be converted to the corresponding Avro types yet.
53	table_name:allcomplextypes, constraint:restrict_to, table_format:text/none/none
54	table_name:allcomplextypes, constraint:restrict_to, table_format:parquet/none/none
55	table_name:allcomplextypes, constraint:restrict_to, table_format:hbase/none/none
56	table_name:functional, constraint:restrict_to, table_format:text/none/none
57	table_name:complextypes_fileformat, constraint:restrict_to, table_format:text/none/none
58	table_name:complextypes_fileformat, constraint:restrict_to, table_format:parquet/none/none
59	table_name:complextypes_fileformat, constraint:restrict_to, table_format:avro/snap/block
60	table_name:complextypes_fileformat, constraint:restrict_to, table_format:rc/snap/block
61	table_name:complextypes_fileformat, constraint:restrict_to, table_format:seq/snap/block
62	table_name:complextypes_multifileformat, constraint:restrict_to, table_format:text/none/none
63	# TODO: Avro
64	table_name:complextypestbl, constraint:restrict_to, table_format:parquet/none/none
65	table_name:alltypeserror, constraint:exclude, table_format:parquet/none/none
66	table_name:alltypeserrornonulls, constraint:exclude, table_format:parquet/none/none
67	table_name:unsupported_types, constraint:exclude, table_format:parquet/none/none
68	table_name:escapechartesttable, constraint:exclude, table_format:parquet/none/none
69	table_name:TblWithRaggedColumns, constraint:exclude, table_format:parquet/none/none
70	# the text_ tables are for testing test delimiters and escape chars in text files
71	table_name:text_comma_backslash_newline, constraint:restrict_to, table_format:text/none/none
72	table_name:text_dollar_hash_pipe, constraint:restrict_to, table_format:text/none/none
73	table_name:text_thorn_ecirc_newline, constraint:restrict_to, table_format:text/none/none
74	table_name:bad_serde, constraint:restrict_to, table_format:text/none/none
75	table_name:rcfile_lazy_binary_serde, constraint:restrict_to, table_format:rc/none/none
76	table_name:unsupported_partition_types, constraint:restrict_to, table_format:text/none/none
77	table_name:nullformat_custom, constraint:exclude, table_format:parquet/none/none
78	table_name:alltypes_view, constraint:restrict_to, table_format:text/none/none
79	table_name:allcomplextypes_view, constraint:restrict_to, table_format:text/none/none
80	table_name:alltypes_view, constraint:restrict_to, table_format:seq/snap/block
81	table_name:alltypes_hive_view, constraint:restrict_to, table_format:text/none/none
82	table_name:alltypes_view_sub, constraint:restrict_to, table_format:text/none/none
83	table_name:alltypes_view_sub, constraint:restrict_to, table_format:seq/snap/block
84	table_name:alltypes_parens, constraint:restrict_to, table_format:text/none/none
85	table_name:complex_view, constraint:restrict_to, table_format:text/none/none
86	table_name:complex_view, constraint:restrict_to, table_format:seq/snap/block
87	table_name:view_view, constraint:restrict_to, table_format:text/none/none
88	table_name:view_view, constraint:restrict_to, table_format:seq/snap/block
89	table_name:subquery_view, constraint:restrict_to, table_format:seq/snap/block
90	table_name:subquery_view, constraint:restrict_to, table_format:rc/none/none
91	# liketbl and tblwithraggedcolumns all have
92	# NULLs in primary key columns. hbase does not support
93	# writing NULLs to primary key columns.
94	table_name:liketbl, constraint:exclude, table_format:hbase/none/none
95	table_name:tblwithraggedcolumns, constraint:exclude, table_format:hbase/none/none
96	# Tables with only one column are not supported in hbase.
97	table_name:greptiny, constraint:exclude, table_format:hbase/none/none
98	table_name:tinyinttable, constraint:exclude, table_format:hbase/none/none
99	# overflow uses a manually constructed text file which doesn't make sense to write to
100	# other table formats since the values that would be written are different (e.g. already
101	# truncated.)
102	table_name:overflow, constraint:restrict_to, table_format:text/none/none
103	# widerow has a single column with a single row containing a 10MB string. hbase doesn't
104	# seem to like this.
105	table_name:widerow, constraint:exclude, table_format:hbase/none/none
106	# nullformat_custom is used in null-insert tests, which user insert overwrite,
107	# which is not supported in hbase. The schema is also specified in HIVE_CREATE
108	# with no corresponding LOAD statement.
109	table_name:nullformat_custom, constraint:exclude, table_format:hbase/none/none
110	table_name:unsupported_types, constraint:exclude, table_format:hbase/none/none
111	# Decimal can only be tested on formats Impala can write to (text and parquet).
112	# TODO: add Avro once Hive or Impala can write Avro decimals
113	table_name:decimal_tbl, constraint:restrict_to, table_format:text/none/none
114	table_name:decimal_tiny, constraint:restrict_to, table_format:text/none/none
115	table_name:decimal_tbl, constraint:restrict_to, table_format:parquet/none/none
116	table_name:decimal_tiny, constraint:restrict_to, table_format:parquet/none/none
117	table_name:avro_decimal_tbl, constraint:restrict_to, table_format:avro/snap/block
118	# TODO first set of tests are for text/none/none
119	table_name:chars_tiny, constraint:restrict_to, table_format:text/none/none
120	# invalid_decimal_part_tbl[1,2,3] tables are used for testing invalid decimal
121	# partition key values (see IMPALA-1040)
122	table_name:invalid_decimal_part_tbl1, constraint:restrict_to, table_format:text/none/none
123	table_name:invalid_decimal_part_tbl2, constraint:restrict_to, table_format:text/none/none
124	table_name:invalid_decimal_part_tbl3, constraint:restrict_to, table_format:text/none/none
125	table_name:avro_decimal_tbl, constraint:restrict_to, table_format:avro/snap/block
126	# testescape tables are used for testing text scanner delimiter handling
127	table_name:table_no_newline, constraint:restrict_to, table_format:text/none/none
128	table_name:table_no_newline_part, constraint:restrict_to, table_format:text/none/none
129	table_name:testescape_16_lf, constraint:restrict_to, table_format:text/none/none
130	table_name:testescape_16_crlf, constraint:restrict_to, table_format:text/none/none
131	table_name:testescape_17_lf, constraint:restrict_to, table_format:text/none/none
132	table_name:testescape_17_crlf, constraint:restrict_to, table_format:text/none/none
133	table_name:testescape_32_lf, constraint:restrict_to, table_format:text/none/none
134	table_name:testescape_32_crlf, constraint:restrict_to, table_format:text/none/none
135	# alltimezones is used to verify that impala properly deals with timezones
136	table_name:alltimezones, constraint:restrict_to, table_format:text/none/none
137	# Avro schema is inferred from the column definitions (IMPALA-1136)
138	table_name:no_avro_schema, constraint:restrict_to, table_format:avro/snap/block
139	table_name:avro_unicode_nulls, constraint:restrict_to, table_format:avro/snap/block
140	# test single and multi stream bz2 files
141	table_name:bzip2_tbl, constraint:restrict_to, table_format:text/bzip/block
142	table_name:large_bzip2_tbl, constraint:restrict_to, table_format:text/bzip/block
143	table_name:multistream_bzip2_tbl, constraint:restrict_to, table_format:text/bzip/block
144	table_name:large_multistream_bzip2_tbl, constraint:restrict_to, table_format:text/bzip/block
145	# Kudu can't handle certain types such as timestamp so we pick and choose the tables
146	# we actually use for Kudu related tests.
147	table_name:alltypes, constraint:only, table_format:kudu/none/none
148	table_name:alltypessmall, constraint:only, table_format:kudu/none/none
149	table_name:alltypestiny, constraint:only, table_format:kudu/none/none
150	table_name:alltypesagg, constraint:only, table_format:kudu/none/none
151	table_name:alltypesaggnonulls, constraint:only, table_format:kudu/none/none
152	table_name:testtbl, constraint:only, table_format:kudu/none/none
153	table_name:jointbl, constraint:only, table_format:kudu/none/none
154	table_name:emptytable, constraint:only, table_format:kudu/none/none
155	table_name:dimtbl, constraint:only, table_format:kudu/none/none
156	table_name:tinytable, constraint:only, table_format:kudu/none/none
157	table_name:tinyinttable, constraint:only, table_format:kudu/none/none
158	table_name:zipcode_incomes, constraint:only, table_format:kudu/none/none
159	table_name:nulltable, constraint:only, table_format:kudu/none/none
160	table_name:nullescapedtable, constraint:only, table_format:kudu/none/none
161	# Skipping header lines is only effective with text tables
162	table_name:table_with_header, constraint:restrict_to, table_format:text/none/none
163	table_name:table_with_header_2, constraint:restrict_to, table_format:text/none/none
164	table_name:table_with_header_insert, constraint:restrict_to, table_format:text/none/none
165	# We also test that skipping header lines works on compressed tables (IMPALA-5287)
166	table_name:table_with_header, constraint:restrict_to, table_format:text/gzip/block
167	table_name:table_with_header_2, constraint:restrict_to, table_format:text/gzip/block
168	table_name:table_with_header_insert, constraint:restrict_to, table_format:text/gzip/block
169	# Inserting into parquet tables should not be affected by the 'skip.header.line.count'
170	# property, so we test parquet format as well.
171	table_name:table_with_header_insert, constraint:restrict_to, table_format:parquet/none/none

13 KiB Raw Blame History

13 KiB

Raw Blame History