jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-01-01 09:00:42 -05:00
Code Issues Projects Releases Wiki Activity
Files
808f9a661afb602aa8bfeba5e77458e088a8457b
impala/testdata/workloads/functional-query/queries/QueryTest/compute-stats.test
Alex Behm ce40134ad0 IMPALA-867: Fail COMPUTE STATS in analysis for Avro tables affected by HIVE-6308. 
Avro tables that were not created with a column-definition list do not have
their columns properly populated in the Metastore backend DB (HIVE-6308).
For such tables COMPUTE STATS and Hive's ANALYZE TABLE cannot succeed.
This patch fails COMPUTE STATS in analysis for such broken Avro tables
and adds tests for Avro tables with mismatched a column-definition list
and Avro schema.

Change-Id: I561ecea944ae2f83d69950b7a1ab9edaa89bdcea
Reviewed-on: http://gerrit.ent.cloudera.com:8080/1892
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: jenkins
Reviewed-on: http://gerrit.ent.cloudera.com:8080/1920
2014-03-14 23:24:55 -07:00

507 lines
14 KiB
Plaintext
Raw Blame History

====
---- QUERY
# test computing stats on a partitioned text table with all types
create table compute_stats_db.alltypes like functional.alltypes;
insert into compute_stats_db.alltypes partition(year, month)
select * from functional.alltypes;
====
---- QUERY
compute stats compute_stats_db.alltypes
---- RESULTS
'Updated 24 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.alltypes
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
2009,1,310,1,'24.56KB','TEXT'
2009,2,280,1,'22.27KB','TEXT'
2009,3,310,1,'24.67KB','TEXT'
2009,4,300,1,'24.06KB','TEXT'
2009,5,310,1,'24.97KB','TEXT'
2009,6,300,1,'24.16KB','TEXT'
2009,7,310,1,'24.97KB','TEXT'
2009,8,310,1,'24.97KB','TEXT'
2009,9,300,1,'24.16KB','TEXT'
2009,10,310,1,'24.97KB','TEXT'
2009,11,300,1,'24.16KB','TEXT'
2009,12,310,1,'24.97KB','TEXT'
2010,1,310,1,'24.97KB','TEXT'
2010,2,280,1,'22.54KB','TEXT'
2010,3,310,1,'24.97KB','TEXT'
2010,4,300,1,'24.16KB','TEXT'
2010,5,310,1,'24.97KB','TEXT'
2010,6,300,1,'24.16KB','TEXT'
2010,7,310,1,'24.97KB','TEXT'
2010,8,310,1,'24.97KB','TEXT'
2010,9,300,1,'24.16KB','TEXT'
2010,10,310,1,'24.97KB','TEXT'
2010,11,300,1,'24.16KB','TEXT'
2010,12,310,1,'24.97KB','TEXT'
Total,,7300,24,'586.84KB',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.alltypes
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',8161,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',10,0,1,1
'smallint_col','SMALLINT',10,0,2,2
'int_col','INT',10,0,4,4
'bigint_col','BIGINT',10,0,8,8
'float_col','FLOAT',10,0,4,4
'double_col','DOUBLE',10,0,8,8
'date_string_col','STRING',666,0,8,8
'string_col','STRING',10,0,1,1
'timestamp_col','TIMESTAMP',5678,0,16,16
'year','INT',2,0,4,4
'month','INT',12,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# test computing stats on an partitioned text table with all types
create table compute_stats_db.alltypesnopart like functional.alltypesnopart;
insert into compute_stats_db.alltypesnopart
select id, bool_col, tinyint_col, smallint_col, int_col, bigint_col, float_col,
double_col, date_string_col, string_col, timestamp_col
from functional.alltypessmall;
====
---- QUERY
compute stats compute_stats_db.alltypesnopart
---- RESULTS
'Updated 1 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.alltypesnopart
---- LABELS
#ROWS, #FILES, SIZE, FORMAT
---- RESULTS
100,3,'7.73KB','TEXT'
---- TYPES
BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.alltypesnopart
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',105,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',10,0,1,1
'smallint_col','SMALLINT',10,0,2,2
'int_col','INT',10,0,4,4
'bigint_col','BIGINT',10,0,8,8
'float_col','FLOAT',10,0,4,4
'double_col','DOUBLE',10,0,8,8
'date_string_col','STRING',12,0,8,8
'string_col','STRING',10,0,1,1
'timestamp_col','TIMESTAMP',101,0,16,16
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# test computing stats on a partitioned parquet table with all types
create table compute_stats_db.alltypes_parquet
like functional_parquet.alltypes;
insert into compute_stats_db.alltypes_parquet partition(year, month)
select * from functional.alltypes;
====
---- QUERY
compute stats compute_stats_db.alltypes_parquet
---- RESULTS
'Updated 24 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.alltypes_parquet
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
2009,1,310,1,regex:.+KB,'PARQUET'
2009,2,280,1,regex:.+KB,'PARQUET'
2009,3,310,1,regex:.+KB,'PARQUET'
2009,4,300,1,regex:.+KB,'PARQUET'
2009,5,310,1,regex:.+KB,'PARQUET'
2009,6,300,1,regex:.+KB,'PARQUET'
2009,7,310,1,regex:.+KB,'PARQUET'
2009,8,310,1,regex:.+KB,'PARQUET'
2009,9,300,1,regex:.+KB,'PARQUET'
2009,10,310,1,regex:.+KB,'PARQUET'
2009,11,300,1,regex:.+KB,'PARQUET'
2009,12,310,1,regex:.+KB,'PARQUET'
2010,1,310,1,regex:.+KB,'PARQUET'
2010,2,280,1,regex:.+KB,'PARQUET'
2010,3,310,1,regex:.+KB,'PARQUET'
2010,4,300,1,regex:.+KB,'PARQUET'
2010,5,310,1,regex:.+KB,'PARQUET'
2010,6,300,1,regex:.+KB,'PARQUET'
2010,7,310,1,regex:.+KB,'PARQUET'
2010,8,310,1,regex:.+KB,'PARQUET'
2010,9,300,1,regex:.+KB,'PARQUET'
2010,10,310,1,regex:.+KB,'PARQUET'
2010,11,300,1,regex:.+KB,'PARQUET'
2010,12,310,1,regex:.+KB,'PARQUET'
Total,,7300,24,regex:.+KB,''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.alltypes_parquet
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',8161,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',10,0,1,1
'smallint_col','SMALLINT',10,0,2,2
'int_col','INT',10,0,4,4
'bigint_col','BIGINT',10,0,8,8
'float_col','FLOAT',10,0,4,4
'double_col','DOUBLE',10,0,8,8
'date_string_col','STRING',666,0,8,8
'string_col','STRING',10,0,1,1
'timestamp_col','TIMESTAMP',5678,0,16,16
'year','INT',2,0,4,4
'month','INT',12,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# test computing stats on an HBase table
create table compute_stats_db.alltypessmall_hbase
like functional_hbase.alltypessmall;
====
---- QUERY
compute stats compute_stats_db.alltypessmall_hbase
---- RESULTS
'Updated 1 partition(s) and 13 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.alltypessmall_hbase
---- LABELS
REGION LOCATION, START ROWKEY, EST. #ROWS, SIZE
---- RESULTS: VERIFY_IS_EQUAL
regex:.+,'',regex:.+,regex:.+KB
regex:.+,'1',regex:.+,regex:.+KB
regex:.+,'3',regex:.+,regex:.+KB
regex:.+,'5',regex:.+,regex:.+KB
regex:.+,'7',regex:.+,regex:.+KB
regex:.+,'9',regex:.+,regex:.+KB
'Total','',regex:.+,regex:.+KB
---- TYPES
STRING, STRING, BIGINT, STRING
====
---- QUERY
show column stats compute_stats_db.alltypessmall_hbase
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',105,0,4,4
'bigint_col','BIGINT',10,0,8,8
'bool_col','BOOLEAN',2,0,1,1
'date_string_col','STRING',12,0,8,8
'double_col','DOUBLE',10,0,8,8
'float_col','FLOAT',10,0,4,4
'int_col','INT',10,0,4,4
'month','INT',4,0,4,4
'smallint_col','SMALLINT',10,0,2,2
'string_col','STRING',10,0,1,1
'timestamp_col','TIMESTAMP',101,0,16,16
'tinyint_col','TINYINT',10,0,1,1
'year','INT',1,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# test computing stats on an binary HBase table
create table compute_stats_db.alltypessmall_hbase_bin
like functional_hbase.alltypessmallbinary;
====
---- QUERY
compute stats compute_stats_db.alltypessmall_hbase_bin
---- RESULTS
'Updated 1 partition(s) and 13 column(s).'
---- TYPES
STRING
====
---- QUERY: VERIFY_IS_EQUAL
show table stats compute_stats_db.alltypessmall_hbase_bin
---- LABELS
REGION LOCATION, START ROWKEY, EST. #ROWS, SIZE
---- RESULTS
regex:.+,'',regex:.+,regex:.+
---- TYPES
STRING, STRING, BIGINT, STRING
====
---- QUERY
show column stats compute_stats_db.alltypessmall_hbase_bin
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',105,0,4,4
'bigint_col','BIGINT',10,0,8,8
'bool_col','BOOLEAN',2,0,1,1
'date_string_col','STRING',12,0,8,8
'double_col','DOUBLE',10,0,8,8
'float_col','FLOAT',10,0,4,4
'int_col','INT',10,0,4,4
'month','INT',4,0,4,4
'smallint_col','SMALLINT',10,0,2,2
'string_col','STRING',10,0,1,1
'timestamp_col','TIMESTAMP',101,0,16,16
'tinyint_col','TINYINT',10,0,1,1
'year','INT',1,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# test computing stats on an empty table
create table compute_stats_db.alltypes_empty like functional_rc_snap.alltypes
====
---- QUERY
compute stats compute_stats_db.alltypes_empty
---- RESULTS
'Updated 0 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.alltypes_empty
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
Total,,0,0,'0B',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.alltypes_empty
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',0,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',0,0,1,1
'smallint_col','SMALLINT',0,0,2,2
'int_col','INT',0,0,4,4
'bigint_col','BIGINT',0,0,8,8
'float_col','FLOAT',0,0,4,4
'double_col','DOUBLE',0,0,8,8
'date_string_col','STRING',0,0,0,0
'string_col','STRING',0,0,0,0
'timestamp_col','TIMESTAMP',0,0,16,16
'year','INT',0,0,4,4
'month','INT',0,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# IMP-1227: Test computing stats on an HBase table that has a
# complex-typed column that Impala does not yet support.
create table compute_stats_db.map_table
like functional_hbase.map_table_hbase;
====
---- QUERY
compute stats compute_stats_db.map_table
---- RESULTS
'Updated 1 partition(s) and 1 column(s).'
---- TYPES
STRING
====
---- QUERY: VERIFY_IS_EQUAL
show table stats compute_stats_db.map_table
---- LABELS
REGION LOCATION, START ROWKEY, EST. #ROWS, SIZE
---- RESULTS
regex:.+,'',regex:.+,regex:.+
---- TYPES
STRING, STRING, BIGINT, STRING
====
---- QUERY
show column stats compute_stats_db.map_table
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'key','STRING',0,0,0,0
'map_col','INVALID_TYPE',-1,-1,-1,-1
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# IMPALA-867: Test computing stats on Avro tables with matching/mismatched
# column definitions and Avro schema. Clone the used tables here.
create table compute_stats_db.avro_alltypes
like functional_avro_snap.alltypes;
create table compute_stats_db.avro_alltypes_extra_coldef
like functional_avro_snap.alltypes_extra_coldef;
create table compute_stats_db.avro_alltypes_missing_coldef
like functional_avro_snap.alltypes_missing_coldef;
create table compute_stats_db.avro_alltypes_type_mismatch
like functional_avro_snap.alltypes_type_mismatch;
====
---- QUERY
# Avro table with matching column definitions and Avro schema
compute stats compute_stats_db.avro_alltypes
---- RESULTS
'Updated 0 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.avro_alltypes
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
Total,,0,0,'0B',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.avro_alltypes
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',0,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','INT',0,0,4,4
'smallint_col','INT',0,0,4,4
'int_col','INT',0,0,4,4
'bigint_col','BIGINT',0,0,8,8
'float_col','FLOAT',0,0,4,4
'double_col','DOUBLE',0,0,8,8
'date_string_col','STRING',0,0,0,0
'string_col','STRING',0,0,0,0
'timestamp_col','STRING',0,0,0,0
'year','INT',0,0,4,4
'month','INT',0,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# Avro table with an extra column definition.
compute stats compute_stats_db.avro_alltypes_extra_coldef
---- RESULTS
'Updated 0 partition(s) and 12 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.avro_alltypes_extra_coldef
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
Total,,0,0,'0B',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.avro_alltypes_extra_coldef
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',0,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',0,0,1,1
'smallint_col','SMALLINT',0,0,2,2
'int_col','INT',0,0,4,4
'bigint_col','BIGINT',0,0,8,8
'float_col','FLOAT',0,0,4,4
'double_col','DOUBLE',0,0,8,8
'date_string_col','STRING',0,0,0,0
'string_col','STRING',0,0,0,0
'timestamp_col','TIMESTAMP',0,0,16,16
'extra_col','STRING',0,0,0,0
'year','INT',0,0,4,4
'month','INT',0,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# Avro table with missing two column definitions.
compute stats compute_stats_db.avro_alltypes_missing_coldef
---- RESULTS
'Updated 0 partition(s) and 9 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.avro_alltypes_missing_coldef
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
Total,,0,0,'0B',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.avro_alltypes_missing_coldef
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',0,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'smallint_col','SMALLINT',0,0,2,2
'int_col','INT',0,0,4,4
'bigint_col','BIGINT',0,0,8,8
'float_col','FLOAT',0,0,4,4
'double_col','DOUBLE',0,0,8,8
'date_string_col','STRING',0,0,0,0
'string_col','STRING',0,0,0,0
'year','INT',0,0,4,4
'month','INT',0,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
---- QUERY
# Avro table with one column definition having a different
# type than the Avro schema (bigint_col is a string).
compute stats compute_stats_db.avro_alltypes_type_mismatch
---- RESULTS
'Updated 0 partition(s) and 11 column(s).'
---- TYPES
STRING
====
---- QUERY
show table stats compute_stats_db.avro_alltypes_type_mismatch
---- LABELS
YEAR, MONTH, #ROWS, #FILES, SIZE, FORMAT
---- RESULTS
Total,,0,0,'0B',''
---- TYPES
INT, INT, BIGINT, BIGINT, STRING, STRING
====
---- QUERY
show column stats compute_stats_db.avro_alltypes_type_mismatch
---- LABELS
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
---- RESULTS
'id','INT',0,0,4,4
'bool_col','BOOLEAN',2,0,1,1
'tinyint_col','TINYINT',0,0,1,1
'smallint_col','SMALLINT',0,0,2,2
'int_col','INT',0,0,4,4
'bigint_col','STRING',0,0,0,0
'float_col','FLOAT',0,0,4,4
'double_col','DOUBLE',0,0,8,8
'date_string_col','STRING',0,0,0,0
'string_col','STRING',0,0,0,0
'timestamp_col','TIMESTAMP',0,0,16,16
'year','INT',0,0,4,4
'month','INT',0,0,4,4
---- TYPES
STRING, STRING, BIGINT, BIGINT, DOUBLE, DOUBLE
====
Reference in New Issue View Git Blame Copy Permalink