mirror of
https://github.com/apache/impala.git
synced 2026-01-05 21:00:54 -05:00
6af7c8fe4a
Because we add 'total' to the last row in SHOW PARTITIONS, we set the partition key columns to be string. At least, that's what the comment said, but we didn't do that in fact. This patch also corrects the column type for max width, which should be INT. Change-Id: I787ab17be27f45107340119017e528c58a3daad3 Reviewed-on: http://gerrit.sjc.cloudera.com:8080/4678 Reviewed-by: Henry Robinson <henry@cloudera.com> Tested-by: jenkins
910 lines
27 KiB
Plaintext
910 lines
27 KiB
Plaintext
====
|
|
---- 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, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'2009','1',310,1,'24.56KB','NOT CACHED','TEXT'
|
|
'2009','2',280,1,'22.27KB','NOT CACHED','TEXT'
|
|
'2009','3',310,1,'24.67KB','NOT CACHED','TEXT'
|
|
'2009','4',300,1,'24.06KB','NOT CACHED','TEXT'
|
|
'2009','5',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','6',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2009','7',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','8',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','9',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2009','10',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','11',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2009','12',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','1',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','2',280,1,'22.54KB','NOT CACHED','TEXT'
|
|
'2010','3',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','4',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','5',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','6',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','7',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','8',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','9',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','10',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','11',300,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','12',310,1,'24.97KB','NOT CACHED','TEXT'
|
|
'Total','',7300,24,'586.84KB','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',8161,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',10,-1,1,1
|
|
'smallint_col','SMALLINT',10,-1,2,2
|
|
'int_col','INT',10,-1,4,4
|
|
'bigint_col','BIGINT',10,-1,8,8
|
|
'float_col','FLOAT',10,-1,4,4
|
|
'double_col','DOUBLE',10,-1,8,8
|
|
'date_string_col','STRING',666,-1,8,8
|
|
'string_col','STRING',10,-1,1,1
|
|
'timestamp_col','TIMESTAMP',5678,-1,16,16
|
|
'year','INT',2,0,4,4
|
|
'month','INT',12,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# drop stats from this table
|
|
drop stats compute_stats_db.alltypes
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'2009','1',-1,1,'24.56KB','NOT CACHED','TEXT'
|
|
'2009','10',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','11',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2009','12',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','2',-1,1,'22.27KB','NOT CACHED','TEXT'
|
|
'2009','3',-1,1,'24.67KB','NOT CACHED','TEXT'
|
|
'2009','4',-1,1,'24.06KB','NOT CACHED','TEXT'
|
|
'2009','5',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','6',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2009','7',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','8',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2009','9',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','1',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','10',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','11',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','12',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','2',-1,1,'22.54KB','NOT CACHED','TEXT'
|
|
'2010','3',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','4',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','5',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','6',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'2010','7',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','8',-1,1,'24.97KB','NOT CACHED','TEXT'
|
|
'2010','9',-1,1,'24.16KB','NOT CACHED','TEXT'
|
|
'Total','',-1,24,'586.84KB','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
# Note - the NDV for partition columns is read from the table metadata.
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',-1,-1,4,4
|
|
'bool_col','BOOLEAN',-1,-1,1,1
|
|
'tinyint_col','TINYINT',-1,-1,1,1
|
|
'smallint_col','SMALLINT',-1,-1,2,2
|
|
'int_col','INT',-1,-1,4,4
|
|
'bigint_col','BIGINT',-1,-1,8,8
|
|
'float_col','FLOAT',-1,-1,4,4
|
|
'double_col','DOUBLE',-1,-1,8,8
|
|
'date_string_col','STRING',-1,-1,-1,-1
|
|
'string_col','STRING',-1,-1,-1,-1
|
|
'timestamp_col','TIMESTAMP',-1,-1,16,16
|
|
'year','INT',2,0,4,4
|
|
'month','INT',12,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# Add partitions with NULL values and check for stats.
|
|
alter table compute_stats_db.alltypes add partition (year=NULL, month=NULL)
|
|
---- RESULTS
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',-1,-1,4,4
|
|
'bool_col','BOOLEAN',-1,-1,1,1
|
|
'tinyint_col','TINYINT',-1,-1,1,1
|
|
'smallint_col','SMALLINT',-1,-1,2,2
|
|
'int_col','INT',-1,-1,4,4
|
|
'bigint_col','BIGINT',-1,-1,8,8
|
|
'float_col','FLOAT',-1,-1,4,4
|
|
'double_col','DOUBLE',-1,-1,8,8
|
|
'date_string_col','STRING',-1,-1,-1,-1
|
|
'string_col','STRING',-1,-1,-1,-1
|
|
'timestamp_col','TIMESTAMP',-1,-1,16,16
|
|
'year','INT',3,1,4,4
|
|
'month','INT',13,1,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
alter table compute_stats_db.alltypes add partition (year=2011, month=NULL)
|
|
---- RESULTS
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',-1,-1,4,4
|
|
'bool_col','BOOLEAN',-1,-1,1,1
|
|
'tinyint_col','TINYINT',-1,-1,1,1
|
|
'smallint_col','SMALLINT',-1,-1,2,2
|
|
'int_col','INT',-1,-1,4,4
|
|
'bigint_col','BIGINT',-1,-1,8,8
|
|
'float_col','FLOAT',-1,-1,4,4
|
|
'double_col','DOUBLE',-1,-1,8,8
|
|
'date_string_col','STRING',-1,-1,-1,-1
|
|
'string_col','STRING',-1,-1,-1,-1
|
|
'timestamp_col','TIMESTAMP',-1,-1,16,16
|
|
'year','INT',4,1,4,4
|
|
'month','INT',13,2,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# Drop the partitions with NULL values and check for stats.
|
|
alter table compute_stats_db.alltypes drop partition (year=NULL, month=NULL)
|
|
---- RESULTS
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',-1,-1,4,4
|
|
'bool_col','BOOLEAN',-1,-1,1,1
|
|
'tinyint_col','TINYINT',-1,-1,1,1
|
|
'smallint_col','SMALLINT',-1,-1,2,2
|
|
'int_col','INT',-1,-1,4,4
|
|
'bigint_col','BIGINT',-1,-1,8,8
|
|
'float_col','FLOAT',-1,-1,4,4
|
|
'double_col','DOUBLE',-1,-1,8,8
|
|
'date_string_col','STRING',-1,-1,-1,-1
|
|
'string_col','STRING',-1,-1,-1,-1
|
|
'timestamp_col','TIMESTAMP',-1,-1,16,16
|
|
'year','INT',3,0,4,4
|
|
'month','INT',13,1,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
alter table compute_stats_db.alltypes drop partition (year=2011, month=NULL)
|
|
---- RESULTS
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',-1,-1,4,4
|
|
'bool_col','BOOLEAN',-1,-1,1,1
|
|
'tinyint_col','TINYINT',-1,-1,1,1
|
|
'smallint_col','SMALLINT',-1,-1,2,2
|
|
'int_col','INT',-1,-1,4,4
|
|
'bigint_col','BIGINT',-1,-1,8,8
|
|
'float_col','FLOAT',-1,-1,4,4
|
|
'double_col','DOUBLE',-1,-1,8,8
|
|
'date_string_col','STRING',-1,-1,-1,-1
|
|
'string_col','STRING',-1,-1,-1,-1
|
|
'timestamp_col','TIMESTAMP',-1,-1,16,16
|
|
'year','INT',2,0,4,4
|
|
'month','INT',12,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# drop stats from this table a second time, should not throw an error.
|
|
drop stats compute_stats_db.alltypes
|
|
====
|
|
---- 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, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
100,3,'7.73KB','NOT CACHED','TEXT'
|
|
---- TYPES
|
|
BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.alltypesnopart
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',105,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',10,-1,1,1
|
|
'smallint_col','SMALLINT',10,-1,2,2
|
|
'int_col','INT',10,-1,4,4
|
|
'bigint_col','BIGINT',10,-1,8,8
|
|
'float_col','FLOAT',10,-1,4,4
|
|
'double_col','DOUBLE',10,-1,8,8
|
|
'date_string_col','STRING',12,-1,8,8
|
|
'string_col','STRING',10,-1,1,1
|
|
'timestamp_col','TIMESTAMP',101,-1,16,16
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'2009','1',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','2',280,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','3',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','4',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','5',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','6',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','7',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','8',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','9',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','10',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','11',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2009','12',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','1',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','2',280,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','3',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','4',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','5',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','6',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','7',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','8',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','9',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','10',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','11',300,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'2010','12',310,1,regex:.+KB,'NOT CACHED','PARQUET'
|
|
'Total','',7300,24,regex:.+KB,'0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, 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,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',10,-1,1,1
|
|
'smallint_col','SMALLINT',10,-1,2,2
|
|
'int_col','INT',10,-1,4,4
|
|
'bigint_col','BIGINT',10,-1,8,8
|
|
'float_col','FLOAT',10,-1,4,4
|
|
'double_col','DOUBLE',10,-1,8,8
|
|
'date_string_col','STRING',666,-1,8,8
|
|
'string_col','STRING',10,-1,1,1
|
|
'timestamp_col','TIMESTAMP',5678,-1,16,16
|
|
'year','INT',2,0,4,4
|
|
'month','INT',12,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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,-1,4,4
|
|
'bigint_col','BIGINT',10,-1,8,8
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'date_string_col','STRING',12,-1,8,8
|
|
'double_col','DOUBLE',10,-1,8,8
|
|
'float_col','FLOAT',10,-1,4,4
|
|
'int_col','INT',10,-1,4,4
|
|
'month','INT',4,-1,4,4
|
|
'smallint_col','SMALLINT',10,-1,2,2
|
|
'string_col','STRING',10,-1,1,1
|
|
'timestamp_col','TIMESTAMP',101,-1,16,16
|
|
'tinyint_col','TINYINT',10,-1,1,1
|
|
'year','INT',1,-1,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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,-1,4,4
|
|
'bigint_col','BIGINT',10,-1,8,8
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'date_string_col','STRING',12,-1,8,8
|
|
'double_col','DOUBLE',10,-1,8,8
|
|
'float_col','FLOAT',10,-1,4,4
|
|
'int_col','INT',10,-1,4,4
|
|
'month','INT',4,-1,4,4
|
|
'smallint_col','SMALLINT',10,-1,2,2
|
|
'string_col','STRING',10,-1,1,1
|
|
'timestamp_col','TIMESTAMP',101,-1,16,16
|
|
'tinyint_col','TINYINT',10,-1,1,1
|
|
'year','INT',1,-1,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, 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,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',0,-1,1,1
|
|
'smallint_col','SMALLINT',0,-1,2,2
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','TIMESTAMP',0,-1,16,16
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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.allcomplextypes
|
|
like functional_hbase.allcomplextypes
|
|
====
|
|
---- QUERY
|
|
compute stats compute_stats_db.allcomplextypes
|
|
---- RESULTS
|
|
'Updated 1 partition(s) and 3 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY: VERIFY_IS_EQUAL
|
|
show table stats compute_stats_db.allcomplextypes
|
|
---- LABELS
|
|
REGION LOCATION, START ROWKEY, EST. #ROWS, SIZE
|
|
---- RESULTS
|
|
regex:.+,'',regex:.+,regex:.+
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.allcomplextypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'array_array_col','ARRAY<ARRAY<INT>>',-1,-1,-1,-1
|
|
'array_map_col','MAP<STRING,ARRAY<INT>>',-1,-1,-1,-1
|
|
'complex_nested_struct_col','STRUCT<f1:INT,f2:ARRAY<STRUCT<f11:BIGINT,f12:MAP<STRING,STRUCT<f21:BIGINT>>>>>',-1,-1,-1,-1
|
|
'complex_struct_col','STRUCT<f1:INT,f2:ARRAY<INT>,f3:MAP<STRING,INT>>',-1,-1,-1,-1
|
|
'int_array_col','ARRAY<INT>',-1,-1,-1,-1
|
|
'int_map_col','MAP<STRING,INT>',-1,-1,-1,-1
|
|
'int_struct_col','STRUCT<f1:INT,f2:INT>',-1,-1,-1,-1
|
|
'map_array_col','ARRAY<MAP<STRING,INT>>',-1,-1,-1,-1
|
|
'map_map_col','MAP<STRING,MAP<STRING,INT>>',-1,-1,-1,-1
|
|
'month','INT',0,-1,4,4
|
|
'nested_struct_col','STRUCT<f1:INT,f2:STRUCT<f11:BIGINT,f12:STRUCT<f21:BIGINT>>>',-1,-1,-1,-1
|
|
'struct_array_col','ARRAY<STRUCT<f1:BIGINT,f2:STRING>>',-1,-1,-1,-1
|
|
'struct_map_col','MAP<STRING,STRUCT<f1:BIGINT,f2:STRING>>',-1,-1,-1,-1
|
|
'year','INT',0,-1,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# IMPALA-867: Test computing stats on Avro tables created by Hive with
|
|
# matching/mismatched column definitions and Avro schema.
|
|
# Clone the used tables here.
|
|
create table compute_stats_db.avro_hive_alltypes
|
|
like functional_avro_snap.alltypes;
|
|
create table compute_stats_db.avro_hive_alltypes_extra_coldef
|
|
like functional_avro_snap.alltypes_extra_coldef;
|
|
create table compute_stats_db.avro_hive_alltypes_missing_coldef
|
|
like functional_avro_snap.alltypes_missing_coldef;
|
|
create table compute_stats_db.avro_hive_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_hive_alltypes
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 11 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_hive_alltypes
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_hive_alltypes
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','INT',0,-1,4,4
|
|
'smallint_col','INT',0,-1,4,4
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# Avro table with an extra column definition.
|
|
compute stats compute_stats_db.avro_hive_alltypes_extra_coldef
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 12 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_hive_alltypes_extra_coldef
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_hive_alltypes_extra_coldef
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',0,-1,1,1
|
|
'smallint_col','SMALLINT',0,-1,2,2
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','TIMESTAMP',0,-1,16,16
|
|
'extra_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# Avro table with missing two column definitions.
|
|
compute stats compute_stats_db.avro_hive_alltypes_missing_coldef
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 9 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_hive_alltypes_missing_coldef
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_hive_alltypes_missing_coldef
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'smallint_col','SMALLINT',0,-1,2,2
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, 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_hive_alltypes_type_mismatch
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 11 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_hive_alltypes_type_mismatch
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_hive_alltypes_type_mismatch
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','TINYINT',0,-1,1,1
|
|
'smallint_col','SMALLINT',0,-1,2,2
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','STRING',0,-1,0,0
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','TIMESTAMP',0,-1,16,16
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# Test Avro table created without any column definitions.
|
|
create table compute_stats_db.avro_impala_alltypes_no_coldefs
|
|
partitioned by (year int, month int)
|
|
with serdeproperties
|
|
('avro.schema.url'='hdfs:///test-warehouse/avro_schemas/functional/alltypes.json')
|
|
stored as avro;
|
|
====
|
|
---- QUERY
|
|
compute stats compute_stats_db.avro_impala_alltypes_no_coldefs
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 11 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_impala_alltypes_no_coldefs
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_impala_alltypes_no_coldefs
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','INT',0,-1,4,4
|
|
'smallint_col','INT',0,-1,4,4
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# IMPALA-1104: Test computing stats on Avro tables created by Impala
|
|
# with mismatched column definitions and Avro schema. Mismatched column name.
|
|
create table compute_stats_db.avro_impala_alltypes_bad_colname
|
|
(id int, bool_col boolean, tinyint_col int, smallint_col int, bad_int_col int,
|
|
bigint_col bigint, float_col float, double_col double, date_string_col string,
|
|
string_col string, timestamp_col timestamp)
|
|
partitioned by (year int, month int)
|
|
with serdeproperties
|
|
('avro.schema.url'='hdfs:///test-warehouse/avro_schemas/functional/alltypes.json')
|
|
stored as avro;
|
|
====
|
|
---- QUERY
|
|
compute stats compute_stats_db.avro_impala_alltypes_bad_colname
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 11 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_impala_alltypes_bad_colname
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_impala_alltypes_bad_colname
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','INT',0,-1,4,4
|
|
'smallint_col','INT',0,-1,4,4
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# IMPALA-1104: Test computing stats on Avro tables created by Impala
|
|
# with mismatched column definitions and Avro schema. Mismatched column type.
|
|
create table compute_stats_db.avro_impala_alltypes_bad_coltype
|
|
(id int, bool_col boolean, tinyint_col int, smallint_col int, int_col int,
|
|
bigint_col bigint, float_col float, double_col bigint, date_string_col string,
|
|
string_col string, timestamp_col timestamp)
|
|
partitioned by (year int, month int)
|
|
with serdeproperties
|
|
('avro.schema.url'='hdfs:///test-warehouse/avro_schemas/functional/alltypes.json')
|
|
stored as avro;
|
|
====
|
|
---- QUERY
|
|
compute stats compute_stats_db.avro_impala_alltypes_bad_coltype
|
|
---- RESULTS
|
|
'Updated 0 partition(s) and 11 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.avro_impala_alltypes_bad_coltype
|
|
---- LABELS
|
|
YEAR, MONTH, #ROWS, #FILES, SIZE, BYTES CACHED, FORMAT
|
|
---- RESULTS
|
|
'Total','',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
show column stats compute_stats_db.avro_impala_alltypes_bad_coltype
|
|
---- LABELS
|
|
COLUMN, TYPE, #DISTINCT VALUES, #NULLS, MAX SIZE, AVG SIZE
|
|
---- RESULTS
|
|
'id','INT',0,-1,4,4
|
|
'bool_col','BOOLEAN',2,-1,1,1
|
|
'tinyint_col','INT',0,-1,4,4
|
|
'smallint_col','INT',0,-1,4,4
|
|
'int_col','INT',0,-1,4,4
|
|
'bigint_col','BIGINT',0,-1,8,8
|
|
'float_col','FLOAT',0,-1,4,4
|
|
'double_col','DOUBLE',0,-1,8,8
|
|
'date_string_col','STRING',0,-1,0,0
|
|
'string_col','STRING',0,-1,0,0
|
|
'timestamp_col','STRING',0,-1,0,0
|
|
'year','INT',0,0,4,4
|
|
'month','INT',0,0,4,4
|
|
---- TYPES
|
|
STRING, STRING, BIGINT, BIGINT, INT, DOUBLE
|
|
====
|
|
---- QUERY
|
|
# For IMPALA-1055, using a database called `parquet` to test cases where the name of
|
|
# the database is a keyword.
|
|
CREATE TABLE `parquet`.billion_parquet(id INT);
|
|
====
|
|
---- QUERY
|
|
COMPUTE STATS `parquet`.billion_parquet
|
|
---- RESULTS
|
|
'Updated 1 partition(s) and 1 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
CREATE TABLE `parquet`.`parquet`(id INT)
|
|
====
|
|
---- QUERY
|
|
COMPUTE STATS `parquet`.`parquet`
|
|
---- RESULTS
|
|
'Updated 1 partition(s) and 1 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
# IMPALA-883: Compute table stats for an empty partition.
|
|
create table compute_stats_db.empty_partitioned (i int) partitioned by (j int);
|
|
alter table compute_stats_db.empty_partitioned add partition (j=1);
|
|
====
|
|
---- QUERY
|
|
compute stats compute_stats_db.empty_partitioned
|
|
---- RESULTS
|
|
'Updated 1 partition(s) and 1 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.empty_partitioned
|
|
---- RESULTS
|
|
'1',0,0,'0B','NOT CACHED','TEXT'
|
|
'Total',0,0,'0B','0B',''
|
|
---- TYPES
|
|
STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
# Insert non empty partition to the table with empty partition.
|
|
insert into table compute_stats_db.empty_partitioned partition (j=2) select 1;
|
|
====
|
|
---- QUERY
|
|
# Verify incremental partition stats work with empty and non-empty partition.
|
|
compute stats compute_stats_db.empty_partitioned
|
|
---- RESULTS
|
|
'Updated 2 partition(s) and 1 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.empty_partitioned
|
|
---- RESULTS
|
|
'1',0,0,'0B','NOT CACHED','TEXT'
|
|
'2',1,1,'2B','NOT CACHED','TEXT'
|
|
'Total',1,1,'2B','0B',''
|
|
---- TYPES
|
|
STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|
|
---- QUERY
|
|
# Verify partition stats work with empty and non-empty partition.
|
|
drop stats compute_stats_db.empty_partitioned;
|
|
compute stats compute_stats_db.empty_partitioned;
|
|
---- RESULTS
|
|
'Updated 2 partition(s) and 1 column(s).'
|
|
---- TYPES
|
|
STRING
|
|
====
|
|
---- QUERY
|
|
show table stats compute_stats_db.empty_partitioned
|
|
---- RESULTS
|
|
'1',0,0,'0B','NOT CACHED','TEXT'
|
|
'2',1,1,'2B','NOT CACHED','TEXT'
|
|
'Total',1,1,'2B','0B',''
|
|
---- TYPES
|
|
STRING, BIGINT, BIGINT, STRING, STRING, STRING
|
|
====
|