jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-01-01 18:00:30 -05:00
Code Issues Projects Releases Wiki Activity
Files
4ad15bb2bec313d5053cd2e07b54c8d904796d66
impala/testdata/workloads/functional-query/queries/QueryTest/create.test
Henry Robinson 44f57e5fb6 IMPALA-1122: Compute stats with partition granularity 
This patch adds the ability to compute and drop column and table
statistics at partition granularity.

The following commands are added. Detail about the implementation
follows.

COMPUTE INCREMENTAL STATS <tbl_name> [PARTITION <partition_spec>]

This variant of COMPUTE STATS will, ultimately, do the same thing as the
traditional COMPUTE STATS statement, but does so by caching the
intermediate state of the computation for each partition in the Hive
MetaStore. If the PARTITION clause is added, the computation is
performed for only that partition. If the PARTITION clause is omitted,
incremental stats are updated only for those partitions with missing
incremental stats (e.g. one column does not have stats, or incremental
stats was never computed for this partition). In this patch, incremental
stats are only invalidated when a DROP STATS variant is executed. Future
patches can automatically invalidate the statistics after REFRESH or
INSERT queries, etc.

DROP INCREMENTAL STATS <tbl_name> PARTITION <part_spec>

This variant of DROP stats removes the incremental statistics for the
given table. It does *not* recalculate the statistics for the whole
table, so this should be used only to invalidate the intermediate state
for a partition which will shortly be subject to COMPUTE INCREMENTAL
STATS. The point of this variant is to allow users to notify Impala when
they believe a partition has changed significantly enough to warrant
recomputation of its statistics. It is not necessary for new partitions;
Impala will detect that they do not have any valid statistics.

--------

This is achieved by adapting the existing HLL UDA via swapping its
finalize method for a new one which returns the intermediate HLL
buckets, rather than aggregating and then disposing of them. This
intermediate state is then returned to Impala's catalog-op-executor.cc,
which then passes the intermediate state back to the frontend to be
ultimately stored in the HMS.

This intermediate state is computed on a per-partition basis by grouping
the input to the UDA by partition. Thus, the incremental computation
produces one row for each partition selected (the set of which might be
quite small, if there are few partitions without valid incremental
stats: this is the point of the new commands).

At the same time, the query coordinator aggregates the output of the UDA
to produce table-level statistics. This computation incorporates any
existing (and not re-computed) intermediate partition state which is
passed to the coordinator by the frontend. The resulting statistics are
saved to the table as normal.

Intermediate statistics are serialised to the HMS by writing a Thrift
structure's serialised form to the partition's 'parameters' map. There
is a schema-imposed limit of 4000 characters to the serialised string,
which is exacerbated by the fact that the Thrift representation must
first be base-64 encoded to avoid type errors in the HMS. The current
patch breaks the encoded structure into 4k chunks, and then recombines
them on read. The alltypes table (11 columns) takes about three of these
chunks. This may mean that incremental stats are not suitable for
particularly wide tables: these structures could be zipped before
encoding for some space savings. In the meantime, the NDV estimates are
run-length encoded (since they are generally sparse); this can result in
substantial space savings.

Change-Id: If82cf4753d19eb532265acb556f798b95fbb0f34
Reviewed-on: http://gerrit.sjc.cloudera.com:8080/4475
Tested-by: jenkins
Reviewed-by: Henry Robinson <henry@cloudera.com>
Reviewed-on: http://gerrit.sjc.cloudera.com:8080/5408
2014-11-25 09:13:37 -08:00

652 lines
15 KiB
Plaintext
Raw Blame History

====
---- QUERY
drop table if exists ddl_test_db.temp_decimal_table
---- RESULTS
====
---- QUERY
create table ddl_test_db.temp_decimal_table like parquet
'/test-warehouse/schemas/decimal.parquet'
---- RESULTS
====
---- QUERY
describe ddl_test_db.temp_decimal_table
---- RESULTS
'd32','decimal(3,2)','inferred from: required fixed_len_byte_array(16) d32 (DECIMAL(3,2))'
'd11','decimal(1,1)','inferred from: required fixed_len_byte_array(16) d11 (DECIMAL(1,1))'
'd1015','decimal(15,10)','inferred from: required fixed_len_byte_array(16) d1015 (DECIMAL(15,10))'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
drop table if exists ddl_test_db.temp_decimal_table
---- RESULTS
====
---- QUERY
create table ddl_test_db.test_like_file_create like parquet
'/test-warehouse/schemas/zipcode_incomes.parquet'
---- RESULTS
====
---- QUERY
describe ddl_test_db.test_like_file_create
---- RESULTS
'id','string','inferred from: optional binary id'
'zip','string','inferred from: optional binary zip'
'description1','string','inferred from: optional binary description1'
'description2','string','inferred from: optional binary description2'
'income','int','inferred from: optional int32 income'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
drop table ddl_test_db.test_like_file_create
---- RESULTS
====
---- QUERY
create table if not exists ddl_test_db.test_like_file_create like parquet
'/test-warehouse/schemas/alltypestiny.parquet'
---- RESULTS
====
---- QUERY
describe ddl_test_db.test_like_file_create
---- RESULTS
'id','int','inferred from: optional int32 id'
'bool_col','boolean','inferred from: optional boolean bool_col'
'tinyint_col','int','inferred from: optional int32 tinyint_col'
'smallint_col','int','inferred from: optional int32 smallint_col'
'int_col','int','inferred from: optional int32 int_col'
'bigint_col','bigint','inferred from: optional int64 bigint_col'
'float_col','float','inferred from: optional float float_col'
'double_col','double','inferred from: optional double double_col'
'date_string_col','string','inferred from: optional binary date_string_col'
'string_col','string','inferred from: optional binary string_col'
'timestamp_col','timestamp','inferred from: optional int96 timestamp_col'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
create table if not exists ddl_test_db.test_like_file_create like parquet
'/test-warehouse/schemas/zipcode_incomes.parquet'
---- RESULTS
====
---- QUERY
# should not have changed since last statement was IF NOT EXISTS
describe ddl_test_db.test_like_file_create
---- RESULTS
'id','int','inferred from: optional int32 id'
'bool_col','boolean','inferred from: optional boolean bool_col'
'tinyint_col','int','inferred from: optional int32 tinyint_col'
'smallint_col','int','inferred from: optional int32 smallint_col'
'int_col','int','inferred from: optional int32 int_col'
'bigint_col','bigint','inferred from: optional int64 bigint_col'
'float_col','float','inferred from: optional float float_col'
'double_col','double','inferred from: optional double double_col'
'date_string_col','string','inferred from: optional binary date_string_col'
'string_col','string','inferred from: optional binary string_col'
'timestamp_col','timestamp','inferred from: optional int96 timestamp_col'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
drop table ddl_test_db.test_like_file_create
---- RESULTS
====
---- QUERY
# It should show up now
show databases like 'ddl_test_db'
---- RESULTS
'ddl_test_db'
---- TYPES
STRING
====
---- QUERY
# Make sure creating a database with the same name doesn't throw an error when
# IF NOT EXISTS is specified.
create database if not exists ddl_test_db
---- RESULTS
====
---- QUERY
show tables in ddl_test_db
---- RESULTS
====
---- QUERY
create table ddl_test_db.testtbl(i int, s string COMMENT 'String col') STORED AS TEXTFILE
---- RESULTS
====
---- QUERY
# Make sure creating a table with the same name doesn't throw an error when
# IF NOT EXISTS is specified.
create table if not exists ddl_test_db.testtbl(i int, s string)
ROW FORMAT DELIMITED
FIELDS TERMINATED BY '\t'
ESCAPED BY '\\'
LINES TERMINATED BY '\n'
STORED AS TEXTFILE
---- RESULTS
====
---- QUERY
show tables in ddl_test_db
---- RESULTS
'testtbl'
---- TYPES
STRING
====
---- QUERY
describe ddl_test_db.testtbl
---- RESULTS
'i','int',''
's','string','String col'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
insert overwrite table ddl_test_db.testtbl SELECT 1, 'Hi'
from functional.alltypes limit 10
---- RESULTS
: 10
====
---- QUERY
select * from ddl_test_db.testtbl
---- RESULTS
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
---- TYPES
INT, STRING
====
---- QUERY
create table ddl_test_db.testtbl_part(i int, s string) PARTITIONED BY (id int comment 'C')
---- RESULTS
====
---- QUERY
# Partition columns are displayed as part of DESCRIBE <table>
describe ddl_test_db.testtbl_part
---- RESULTS
'i','int',''
's','string',''
'id','int','C'
---- TYPES
STRING, STRING, STRING
====
---- QUERY
insert overwrite table ddl_test_db.testtbl_part partition(id=1)
select 10, 'Ten' from functional.alltypes limit 1
---- RESULTS
id=1/: 1
====
---- QUERY
insert overwrite table ddl_test_db.testtbl_part partition(id=2)
select 20, 'Twenty' from functional.alltypes limit 2
---- RESULTS
id=2/: 2
====
---- QUERY
select * from ddl_test_db.testtbl_part
---- RESULTS
10,'Ten',1
20,'Twenty',2
20,'Twenty',2
---- TYPES
INT, STRING, INT
====
---- QUERY
select * from ddl_test_db.testtbl_part where id = 1
---- RESULTS
10,'Ten',1
---- TYPES
INT, STRING, INT
====
---- QUERY
use ddl_test_db
---- RESULTS
====
---- QUERY
show tables
---- RESULTS
'testtbl'
'testtbl_part'
---- TYPES
STRING
====
---- QUERY
# Make sure we create the table in the proper database after a "use"
create table testtbl2(f float, d double) ROW FORMAT DELIMITED FIELDS TERMINATED BY '|'
---- RESULTS
====
---- QUERY
show tables
---- RESULTS
'testtbl'
'testtbl2'
'testtbl_part'
---- TYPES
STRING
====
---- QUERY
drop table testtbl2
---- RESULTS
====
---- QUERY
show tables
---- RESULTS
'testtbl'
'testtbl_part'
---- TYPES
STRING
====
---- QUERY
# CREATE TABLE LIKE on partitioned table
create table alltypes_test like functional_seq_snap.alltypes
stored as parquet
---- RESULTS
====
---- QUERY
# CREATE TABLE LIKE on a view
create table like_view like functional.view_view
---- RESULTS
====
---- QUERY
describe like_view
---- RESULTS
'id','int',''
'bool_col','boolean',''
'tinyint_col','tinyint',''
'smallint_col','smallint',''
'int_col','int',''
'bigint_col','bigint',''
'float_col','float',''
'double_col','double',''
'date_string_col','string',''
'string_col','string',''
'timestamp_col','timestamp',''
'year','int',''
'month','int',''
---- TYPES
STRING, STRING, STRING
====
---- QUERY
show table stats like_view
---- LABELS
#ROWS, #FILES, SIZE, BYTES CACHED, FORMAT, INCREMENTAL STATS
---- RESULTS
-1,0,'0B','NOT CACHED','TEXT','false'
---- TYPES
BIGINT, BIGINT, STRING, STRING, STRING, STRING
====
---- QUERY
drop table like_view
---- RESULTS
====
---- QUERY
create table like_view_parquet like functional.view_view stored as parquet
---- RESULTS
====
---- QUERY
show table stats like_view_parquet
---- LABELS
#ROWS, #FILES, SIZE, BYTES CACHED, FORMAT, INCREMENTAL STATS
---- RESULTS
-1,0,'0B','NOT CACHED','PARQUET','false'
---- TYPES
BIGINT, BIGINT, STRING, STRING, STRING, STRING
====
---- QUERY
drop table like_view_parquet
---- RESULTS
====
---- QUERY
# Make sure no data exists for this table
select count(*) from alltypes_test
---- RESULTS
0
---- TYPES
BIGINT
====
---- QUERY
# Should be able to insert into this table
insert overwrite table alltypes_test
partition (year=2009, month=4)
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.alltypes where year=2009 and month=4
---- RESULTS
year=2009/month=4/: 300
====
---- QUERY
# Make sure we can read the new data.
select count(*) from alltypes_test
---- RESULTS
300
---- TYPES
BIGINT
====
---- QUERY
# This should copy the file format from the source table (rc)
create external table jointbl_like like functional_rc_gzip.jointbl
location '/test-warehouse/jointbl_rc_gzip'
---- RESULTS
====
---- QUERY
# should get some results back
select * from jointbl_like order by test_id limit 3
---- RESULTS
1001,'Name1',94611,5000
1002,'Name2',94611,5000
1003,'Name3',94611,5000
---- TYPES
BIGINT, STRING, INT, INT
====
---- QUERY
# CREATE TABLE LIKE on unpartitioned table.
create table testtbl_like like testtbl
---- RESULTS
====
---- QUERY
# Make sure the new table can be queried and no data exists for this table.
select count(*) from testtbl_like
---- RESULTS
0
---- TYPES
BIGINT
====
---- QUERY
# No error is thrown when IF NOT EXISTS is specified and the table already exists.
create table if not exists testtbl_like like testtbl
---- RESULTS
====
---- QUERY
# IF NOT EXISTS also applies when the src table is the same as the new table.
create table if not exists testtbl_like like testtbl_like
---- RESULTS
====
---- QUERY
insert overwrite table testtbl_like
select * from testtbl limit 5
---- RESULTS
: 5
====
---- QUERY
# Make sure we can read the data.
select * from testtbl_like
---- RESULTS
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
1,'Hi'
---- TYPES
INT, STRING
====
---- QUERY
# Ensure that a table can be created using CTAS
create table ctas_join stored as parquet as
select j.*, a.int_col, 1*2
from functional.jointbl j join functional_seq_snap.alltypes a
on (j.alltypes_id=a.id)
---- RESULTS
'Inserted 12 row(s)'
---- TYPES
STRING
====
---- QUERY
describe ctas_join
---- RESULTS
'test_id','bigint',''
'test_name','string',''
'test_zip','int',''
'alltypes_id','int',''
'int_col','int',''
'_c2','smallint',''
---- TYPES
STRING, STRING, STRING
====
---- QUERY
select * from ctas_join
---- RESULTS
1001,'Name1',94611,5000,0,2
1002,'Name2',94611,5000,0,2
1003,'Name3',94611,5000,0,2
1004,'Name4',94611,5000,0,2
1005,'Name5',94611,5000,0,2
1106,'Name6',94612,5000,0,2
1006,'Name16',94612,5000,0,2
1006,'Name6',94616,5000,0,2
1106,'Name16',94612,5000,0,2
1106,'Name6',94616,5000,0,2
1006,'Name16',94616,5000,0,2
1106,'Name16',94616,5000,0,2
---- TYPES
BIGINT, STRING, INT, INT, INT, SMALLINT
====
---- QUERY
# Since the table already exists, the second CTAS should be a no-op
create table if not exists ctas_join stored as parquet as
select j.*, a.int_col, 1*2
from functional.jointbl j join functional_seq_snap.alltypes a
on (j.alltypes_id=a.id) limit 1
---- RESULTS
'Inserted 0 row(s)'
---- TYPES
STRING
====
---- QUERY
select * from ctas_join
---- RESULTS
1001,'Name1',94611,5000,0,2
1002,'Name2',94611,5000,0,2
1003,'Name3',94611,5000,0,2
1004,'Name4',94611,5000,0,2
1005,'Name5',94611,5000,0,2
1106,'Name6',94612,5000,0,2
1006,'Name16',94612,5000,0,2
1006,'Name6',94616,5000,0,2
1106,'Name16',94612,5000,0,2
1106,'Name6',94616,5000,0,2
1006,'Name16',94616,5000,0,2
1106,'Name16',94616,5000,0,2
---- TYPES
BIGINT, STRING, INT, INT, INT, SMALLINT
====
---- QUERY
# Validate CTAS with LIMIT 0
create table if not exists ctas_join_limit0 stored as textfile as
select * from functional.jointbl limit 0
---- RESULTS
'Inserted 0 row(s)'
---- TYPES
STRING
====
---- QUERY
describe ctas_join_limit0
---- RESULTS
'test_id','bigint',''
'test_name','string',''
'test_zip','int',''
'alltypes_id','int',''
---- TYPES
STRING, STRING, STRING
====
---- QUERY
select * from ctas_join_limit0
---- RESULTS
---- TYPES
BIGINT, STRING, INT, INT
====
---- QUERY
# Validate CTAS with LIMIT 0
create table if not exists ctas_join_limit0 stored as textfile as
select * from functional.jointbl limit 0
---- RESULTS
'Inserted 0 row(s)'
---- TYPES
STRING
====
---- QUERY
select * from ctas_join_limit0
---- RESULTS
---- TYPES
BIGINT, STRING, INT, INT
====
---- QUERY
create table if not exists ctas_join_limit0 stored as textfile as
select * from functional.jointbl limit 4
---- RESULTS
'Inserted 0 row(s)'
---- TYPES
STRING
====
---- QUERY
select * from ctas_join_limit0
---- RESULTS
---- TYPES
BIGINT, STRING, INT, INT
====
---- QUERY
create table allcomplextypes_clone like functional.allcomplextypes
stored as parquet
---- RESULTS
====
---- QUERY
describe allcomplextypes_clone
---- TYPES
STRING, STRING, STRING
---- RESULTS
'id','int',''
'int_array_col','array<int>',''
'array_array_col','array<array<int>>',''
'map_array_col','array<map<string,int>>',''
'struct_array_col','array<struct<f1:bigint,f2:string>>',''
'int_map_col','map<string,int>',''
'array_map_col','map<string,array<int>>',''
'map_map_col','map<string,map<string,int>>',''
'struct_map_col','map<string,struct<f1:bigint,f2:string>>',''
'int_struct_col','struct<f1:int,f2:int>',''
'complex_struct_col','struct<f1:int,f2:array<int>,f3:map<string,int>>',''
'nested_struct_col','struct<f1:int,f2:struct<f11:bigint,f12:struct<f21:bigint>>>',''
'complex_nested_struct_col','struct<f1:int,f2:array<struct<f11:bigint,f12:map<string,struct<f21:bigint>>>>>',''
'year','int',''
'month','int',''
====
---- QUERY
# Test creating an unpartitioned Avro table without column definitions.
create table avro_alltypes_nopart
with serdeproperties
('avro.schema.url'='hdfs:///test-warehouse/avro_schemas/functional/alltypes.json')
stored as avro
---- RESULTS
====
---- QUERY
describe avro_alltypes_nopart
---- TYPES
STRING, STRING, STRING
---- RESULTS
'bigint_col','bigint','from deserializer'
'bool_col','boolean','from deserializer'
'date_string_col','string','from deserializer'
'double_col','double','from deserializer'
'float_col','float','from deserializer'
'id','int','from deserializer'
'int_col','int','from deserializer'
'smallint_col','int','from deserializer'
'string_col','string','from deserializer'
'timestamp_col','string','from deserializer'
'tinyint_col','int','from deserializer'
====
---- QUERY
# Test creating a partitioned Avro table without column definitions.
create table avro_alltypes_part
partitioned by (year int, month int)
with serdeproperties
('avro.schema.url'='hdfs:///test-warehouse/avro_schemas/functional/alltypes.json')
stored as avro
---- RESULTS
====
---- QUERY
describe avro_alltypes_part
---- TYPES
STRING, STRING, STRING
---- RESULTS
'bigint_col','bigint','from deserializer'
'bool_col','boolean','from deserializer'
'date_string_col','string','from deserializer'
'double_col','double','from deserializer'
'float_col','float','from deserializer'
'id','int','from deserializer'
'int_col','int','from deserializer'
'month','int',''
'smallint_col','int','from deserializer'
'string_col','string','from deserializer'
'timestamp_col','string','from deserializer'
'tinyint_col','int','from deserializer'
'year','int',''
====
---- QUERY
drop table avro_alltypes_part
---- RESULTS
====
---- QUERY
drop table avro_alltypes_nopart
---- RESULTS
====
---- QUERY
drop table allcomplextypes_clone
---- RESULTS
====
---- QUERY
drop table ctas_join
---- RESULTS
====
---- QUERY
drop table ctas_join_limit0
---- RESULTS
====
---- QUERY
drop table testtbl
---- RESULTS
====
---- QUERY
drop table testtbl_part
---- RESULTS
====
---- QUERY
drop table alltypes_test
---- RESULTS
====
---- QUERY
drop table jointbl_like
---- RESULTS
====
---- QUERY
drop table testtbl_like
---- RESULTS
====
---- QUERY
show tables
---- RESULTS
---- TYPES
STRING
====
---- QUERY
drop table if exists non_existent_db.tbl
---- RESULTS
====
---- QUERY
# Need to switch databases before dropping
use default;
drop database ddl_test_db
---- RESULTS
====
---- QUERY
# Should be gone now
show databases like 'ddl_test_db'
---- RESULTS
---- TYPES
STRING
====
Reference in New Issue View Git Blame Copy Permalink