mirror of
https://github.com/apache/impala.git
synced 2026-02-01 21:00:29 -05:00
c925807b1a
- serialize using bytes instead of stream - avoid unnecessary constructor during deserialization - simplified code slightly - added original exception message to re-thrown generic message Change-Id: I306a2489dac0f4d2d475e8f9987cd58bf95474bb Reviewed-on: http://gerrit.cloudera.org:8080/17818 Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com> Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
562 lines
19 KiB
Plaintext
562 lines
19 KiB
Plaintext
====
|
|
---- QUERY
|
|
# Use a small table for testing Datasketches Theta functions through Impala to make sure
|
|
# that these approximate functions give the correct result. For testing Impala
|
|
# functionality no need to test how Datasketches Theta approximates count distint values
|
|
# so a small table is enough.
|
|
select
|
|
ds_theta_estimate(ds_theta_sketch(tinyint_col)),
|
|
ds_theta_estimate(ds_theta_sketch(int_col)),
|
|
ds_theta_estimate(ds_theta_sketch(bigint_col)),
|
|
ds_theta_estimate(ds_theta_sketch(float_col)),
|
|
ds_theta_estimate(ds_theta_sketch(double_col)),
|
|
ds_theta_estimate(ds_theta_sketch(string_col))
|
|
from functional_parquet.alltypessmall;
|
|
---- RESULTS
|
|
10,10,10,10,10,10
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
select
|
|
ds_theta_sketch_and_estimate(tinyint_col),
|
|
ds_theta_sketch_and_estimate(int_col),
|
|
ds_theta_sketch_and_estimate(bigint_col),
|
|
ds_theta_sketch_and_estimate(float_col),
|
|
ds_theta_sketch_and_estimate(double_col),
|
|
ds_theta_sketch_and_estimate(string_col)
|
|
from functional_parquet.alltypessmall;
|
|
---- RESULTS
|
|
10,10,10,10,10,10
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
# Check that unsupported types give error with ds_theta_sketch().
|
|
select ds_theta_sketch(bool_col) from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch(BOOLEAN)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch(smallint_col) from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch(SMALLINT)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch(cast(date_string_col as date format 'MM/DD/YYYY'))
|
|
from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch(DATE)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch(d1) from functional_parquet.decimal_tbl;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch(DECIMAL(9,0))
|
|
====
|
|
---- QUERY
|
|
# Check that unsupported types give error with ds_theta_sketch_and_estimate().
|
|
select ds_theta_sketch_and_estimate(bool_col) from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch_and_estimate(BOOLEAN)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch_and_estimate(smallint_col) from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch_and_estimate(SMALLINT)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch_and_estimate(cast(date_string_col as date format 'MM/DD/YYYY'))
|
|
from functional_parquet.alltypessmall;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch_and_estimate(DATE)
|
|
====
|
|
---- QUERY
|
|
select ds_theta_sketch_and_estimate(d1) from functional_parquet.decimal_tbl;
|
|
---- CATCH
|
|
AnalysisException: No matching function with signature: ds_theta_sketch_and_estimate(DECIMAL(9,0))
|
|
====
|
|
---- QUERY
|
|
# Check if Theta works with null values.
|
|
select
|
|
ds_theta_estimate(ds_theta_sketch(null_str)),
|
|
ds_theta_estimate(ds_theta_sketch(null_int)),
|
|
ds_theta_estimate(ds_theta_sketch(null_double)),
|
|
ds_theta_estimate(ds_theta_sketch(some_nulls)),
|
|
ds_theta_sketch_and_estimate(null_str),
|
|
ds_theta_sketch_and_estimate(null_int),
|
|
ds_theta_sketch_and_estimate(null_double),
|
|
ds_theta_sketch_and_estimate(some_nulls)
|
|
from functional_parquet.nullrows;
|
|
---- RESULTS
|
|
0,0,0,6,0,0,0,6
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
# Check if Theta works for empty datasets.
|
|
select
|
|
ds_theta_estimate(ds_theta_sketch(field)),
|
|
ds_theta_estimate(ds_theta_sketch(f2)),
|
|
ds_theta_sketch_and_estimate(field),
|
|
ds_theta_sketch_and_estimate(f2)
|
|
from functional_parquet.emptytable;
|
|
---- RESULTS
|
|
0,0,0,0
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
# Write sketches to a table as string and get an estimate from the written sketch.
|
|
# Note, the plan is to write sketches as binary instead of strings. For this we have to
|
|
# wait for the binary support (IMPALA-9482).
|
|
create table sketch_store
|
|
(year int, month int, date_sketch string, float_sketch string)
|
|
stored as parquet;
|
|
insert into sketch_store
|
|
select
|
|
year,
|
|
month,
|
|
ds_theta_sketch(date_string_col),
|
|
ds_theta_sketch(float_col)
|
|
from functional_parquet.alltypessmall
|
|
group by year, month;
|
|
select
|
|
year,
|
|
month,
|
|
ds_theta_estimate(date_sketch),
|
|
ds_theta_estimate(float_sketch)
|
|
from sketch_store order by month;
|
|
---- RESULTS
|
|
2009,1,3,10
|
|
2009,2,3,10
|
|
2009,3,3,10
|
|
2009,4,3,10
|
|
---- TYPES
|
|
INT,INT,BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_estimate returns error for strings that are not serialized sketches.
|
|
select ds_theta_estimate(date_string_col) from functional_parquet.alltypestiny;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_estimate returns error for HLL serialized sketches.
|
|
create table hll_sketch_store (date_sketch string) stored as parquet;
|
|
insert into hll_sketch_store
|
|
select ds_hll_sketch(date_string_col)
|
|
from functional_parquet.alltypessmall;
|
|
select ds_theta_estimate(date_sketch) from hll_sketch_store;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 7
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_estimate returns null for null and empty string inputs.
|
|
select ds_theta_estimate(b), ds_theta_estimate(c) from functional_parquet.nulltable;
|
|
---- RESULTS
|
|
0,0
|
|
---- TYPES
|
|
BIGINT,BIGINT
|
|
====
|
|
---- QUERY
|
|
# Check that sketches made by Hive can be read and used for estimating by Impala.
|
|
select
|
|
ds_theta_estimate(ti) as ti,
|
|
ds_theta_estimate(i) as i,
|
|
ds_theta_estimate(bi) as bi,
|
|
ds_theta_estimate(f) as f,
|
|
ds_theta_estimate(d) as d,
|
|
ds_theta_estimate(s) as s,
|
|
ds_theta_estimate(c) as c,
|
|
ds_theta_estimate(v) as v,
|
|
ds_theta_estimate(nc) as nc
|
|
from theta_sketches_from_hive;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# Unions the sketches from sketch_store and checks if the union produces the same result
|
|
# as if the whole data was sketched together into a single sketch.
|
|
select
|
|
ds_theta_estimate(ds_theta_union(date_sketch)),
|
|
ds_theta_estimate(ds_theta_union(float_sketch))
|
|
from sketch_store;
|
|
---- TYPES
|
|
BIGINT,BIGINT
|
|
---- RESULTS
|
|
12,10
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_union() produces NULL for an empty dataset.
|
|
select ds_theta_union(field) from functional_parquet.emptytable;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_union() produces NULL for NULL inputs.
|
|
select ds_theta_union(null_str) from functional_parquet.nullrows;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# ds_theta_union() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_union(date_string_col) from functional_parquet.alltypestiny where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Get the same sketches from Impala and Hive and put them into the same table. When we
|
|
# get the estimates from the unions of these sketches the expectation is to get the same
|
|
# results as if these sketches were used separately to get the estimates.
|
|
create table theta_sketches_impala_hive like theta_sketches_from_impala stored as parquet;
|
|
insert into theta_sketches_impala_hive select * from theta_sketches_from_hive;
|
|
insert into theta_sketches_impala_hive select * from theta_sketches_from_impala;
|
|
select
|
|
ds_theta_estimate(ds_theta_union(ti)) as ti,
|
|
ds_theta_estimate(ds_theta_union(i)) as i,
|
|
ds_theta_estimate(ds_theta_union(bi)) as bi,
|
|
ds_theta_estimate(ds_theta_union(f)) as f,
|
|
ds_theta_estimate(ds_theta_union(d)) as d,
|
|
ds_theta_estimate(ds_theta_union(s)) as s,
|
|
ds_theta_estimate(ds_theta_union(c)) as c,
|
|
ds_theta_estimate(ds_theta_union(v)) as v,
|
|
ds_theta_estimate(ds_theta_union(nc)) as nc
|
|
from theta_sketches_impala_hive;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# Intersect the sketches in sketch_store and check.
|
|
select
|
|
ds_theta_estimate(ds_theta_intersect(date_sketch)),
|
|
ds_theta_estimate(ds_theta_intersect(float_sketch))
|
|
from sketch_store;
|
|
---- TYPES
|
|
BIGINT,BIGINT
|
|
---- RESULTS
|
|
0,10
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_intersect() produces NULL for an empty dataset.
|
|
select ds_theta_intersect(field) from functional_parquet.emptytable;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_intersect() produces NULL for NULL inputs.
|
|
select ds_theta_intersect(null_str) from functional_parquet.nullrows;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# ds_theta_intersect() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_intersect(date_string_col) from functional_parquet.alltypestiny
|
|
where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Check that sketches made by Hive for intersection and estimation, which is consistent
|
|
# with direct estimation of these sketches.
|
|
select
|
|
ds_theta_estimate(ds_theta_intersect(ti)) as ti,
|
|
ds_theta_estimate(ds_theta_intersect(i)) as i,
|
|
ds_theta_estimate(ds_theta_intersect(bi)) as bi,
|
|
ds_theta_estimate(ds_theta_intersect(f)) as f,
|
|
ds_theta_estimate(ds_theta_intersect(d)) as d,
|
|
ds_theta_estimate(ds_theta_intersect(s)) as s,
|
|
ds_theta_estimate(ds_theta_intersect(c)) as c,
|
|
ds_theta_estimate(ds_theta_intersect(v)) as v,
|
|
ds_theta_estimate(ds_theta_intersect(nc)) as nc
|
|
from theta_sketches_from_hive;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# Intersect the sketches from theta_sketches_impala_hive and checks if the intersection
|
|
# produces the same result as if these sketches were used separately to get the estimates.
|
|
select
|
|
ds_theta_estimate(ds_theta_intersect(ti)) as ti,
|
|
ds_theta_estimate(ds_theta_intersect(i)) as i,
|
|
ds_theta_estimate(ds_theta_intersect(bi)) as bi,
|
|
ds_theta_estimate(ds_theta_intersect(f)) as f,
|
|
ds_theta_estimate(ds_theta_intersect(d)) as d,
|
|
ds_theta_estimate(ds_theta_intersect(s)) as s,
|
|
ds_theta_estimate(ds_theta_intersect(c)) as c,
|
|
ds_theta_estimate(ds_theta_intersect(v)) as v,
|
|
ds_theta_estimate(ds_theta_intersect(nc)) as nc
|
|
from theta_sketches_impala_hive;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# A and B refer to the two input sketches in the order A-not-B.
|
|
# Checks that ds_theta_exclude() returns an null sketch for A, B is NULL inputs.
|
|
select ds_theta_exclude(null_str, some_nulls) from functional_parquet.nullrows
|
|
where id='b';
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_exclude() returns an null sketch for A is NULL inputs.
|
|
select ds_theta_exclude(null_str, sketch) from (
|
|
select null null_str, ds_theta_sketch(some_nulls) sketch from functional_parquet.nullrows
|
|
where id='a'
|
|
) t;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_exclude() returns an empty sketch When A is empty and B is null.
|
|
select ds_theta_estimate(ds_theta_exclude(ds_theta_sketch(f2), null))
|
|
from functional_parquet.emptytable;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_exclude() returns an null sketch When A is null and B is empty.
|
|
select ds_theta_exclude(null, ds_theta_sketch(f2)) from functional_parquet.emptytable;
|
|
---- TYPES
|
|
STRING
|
|
---- RESULTS
|
|
'NULL'
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_exclude() returns an empty sketch When A and B are both empty.
|
|
select ds_theta_estimate(ds_theta_exclude(ds_theta_sketch(field), ds_theta_sketch(f2)))
|
|
from functional_parquet.emptytable;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# ds_theta_exclude() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_exclude(date_string_col, date_string_col) from functional_parquet.alltypestiny
|
|
where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Get the same sketches from Impala and Hive and put them into the same table (different
|
|
# column prefix). The estimate in sketch A but not in sketch B is 0.
|
|
create table theta_sketches_impala_hive2 (
|
|
i_ti string,i_i string,i_bi string,i_f string,i_d string,i_s string,i_c string,i_v string,
|
|
i_nc string,
|
|
h_ti string,h_i string,h_bi string,h_f string,h_d string,h_s string,h_c string,h_v string,
|
|
h_nc string) stored as parquet;
|
|
insert overwrite theta_sketches_impala_hive2 select
|
|
i.ti i_ti, i.i i_i, i.bi i_bi, i.f i_f, i.d i_d, i.s i_s, i.c i_c, i.v i_v,i.nc i_nc,
|
|
h.ti h_ti, h.i h_i, h.bi h_bi, h.f h_f, h.d h_d, h.s h_s, h.c h_c, h.v h_v,h.nc h_nc
|
|
from theta_sketches_from_impala i, theta_sketches_from_hive h;
|
|
select
|
|
ds_theta_estimate(ds_theta_exclude(i_ti, h_ti)) as ti,
|
|
ds_theta_estimate(ds_theta_exclude(i_i, h_i)) as i,
|
|
ds_theta_estimate(ds_theta_exclude(i_bi, h_bi)) as bi,
|
|
ds_theta_estimate(ds_theta_exclude(i_f, h_f)) as f,
|
|
ds_theta_estimate(ds_theta_exclude(i_d, h_d)) as d,
|
|
ds_theta_estimate(ds_theta_exclude(i_s, h_s)) as s,
|
|
ds_theta_estimate(ds_theta_exclude(i_c, h_c)) as c,
|
|
ds_theta_estimate(ds_theta_exclude(i_v, h_v)) as v,
|
|
ds_theta_estimate(ds_theta_exclude(i_nc, h_nc)) as nc
|
|
from theta_sketches_impala_hive2;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
0,0,0,0,0,0,0,0,0
|
|
====
|
|
---- QUERY
|
|
# Check that the result of an a-not-b is a non-empty sketch.
|
|
create table sketch_input (id1 int, id2 int);
|
|
insert into table sketch_input values
|
|
(1, 2), (2, 4), (3, 6), (4, 8), (5, 10), (6, 12), (7, 14), (8, 16), (9, 18), (10, 20);
|
|
create table sketch_intermediate (sketch1 string, sketch2 string) stored as parquet;
|
|
insert overwrite sketch_intermediate
|
|
select ds_theta_sketch(id1), ds_theta_sketch(id2) from sketch_input;
|
|
select ds_theta_estimate(ds_theta_exclude(sketch1, sketch2)) from sketch_intermediate;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
5
|
|
====
|
|
---- QUERY
|
|
# Checks that unioning a valid sketch with a null value result the valid sketch being
|
|
# returned.
|
|
select
|
|
ds_theta_estimate(ds_theta_union_f(date_sketch, null)),
|
|
ds_theta_estimate(ds_theta_union_f(null, float_sketch))
|
|
from sketch_store;
|
|
---- TYPES
|
|
BIGINT,BIGINT
|
|
---- RESULTS
|
|
3,10
|
|
3,10
|
|
3,10
|
|
3,10
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_union_f() returns an empty sketch for an empty sketch.
|
|
select ds_theta_estimate(ds_theta_union_f(ds_theta_sketch(cast(f2 as float)), null))
|
|
from functional_parquet.emptytable;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_union_f() returns an empty sketch for NULL inputs.
|
|
select ds_theta_estimate(ds_theta_union_f(null_str, some_nulls)) from
|
|
functional_parquet.nullrows where id='b';
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# ds_theta_union_f() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_union_f(null, date_string_col) from functional_parquet.alltypestiny
|
|
where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# ds_theta_union_f() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_union_f(date_string_col, null) from functional_parquet.alltypestiny
|
|
where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Unions the sketches from theta_sketches_impala_hive2 and checks if the union produces
|
|
# the same result as if these sketches were used separately to get the estimates.
|
|
select
|
|
ds_theta_estimate(ds_theta_union_f(i_ti, h_ti)) as ti,
|
|
ds_theta_estimate(ds_theta_union_f(i_i, h_i)) as i,
|
|
ds_theta_estimate(ds_theta_union_f(i_bi, h_bi)) as bi,
|
|
ds_theta_estimate(ds_theta_union_f(i_f, h_f)) as f,
|
|
ds_theta_estimate(ds_theta_union_f(i_d, h_d)) as d,
|
|
ds_theta_estimate(ds_theta_union_f(i_s, h_s)) as s,
|
|
ds_theta_estimate(ds_theta_union_f(i_c, h_c)) as c,
|
|
ds_theta_estimate(ds_theta_union_f(i_v, h_v)) as v,
|
|
ds_theta_estimate(ds_theta_union_f(i_nc, h_nc)) as nc
|
|
from theta_sketches_impala_hive2;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# Union two sketches from different columns of sketch_intermediate.
|
|
select ds_theta_estimate(ds_theta_union_f(sketch1, sketch2)) from sketch_intermediate;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
15
|
|
====
|
|
---- QUERY
|
|
# Check if the intersection of a valid sketch with a null value returns an null sketch.
|
|
select
|
|
ds_theta_estimate(ds_theta_intersect_f(date_sketch, null)),
|
|
ds_theta_estimate(ds_theta_intersect_f(null, float_sketch))
|
|
from sketch_store;
|
|
---- TYPES
|
|
BIGINT,BIGINT
|
|
---- RESULTS
|
|
0,0
|
|
0,0
|
|
0,0
|
|
0,0
|
|
====
|
|
---- QUERY
|
|
# Check that ds_theta_intersect_f() returns an empty sketch for 2 empty sketch.
|
|
select ds_theta_estimate(ds_theta_intersect_f(
|
|
ds_theta_sketch(cast(f2 as float)), ds_theta_sketch(cast(f2 as float))))
|
|
from functional_parquet.emptytable;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# Checks that ds_theta_intersect_f() returns an null sketch for 2 null inputs.
|
|
select ds_theta_estimate(ds_theta_intersect_f(null_str, some_nulls)) from
|
|
functional_parquet.nullrows where id='b';
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# ds_theta_intersect_f() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_intersect_f(date_string_col, null) from functional_parquet.alltypestiny
|
|
where id=1;
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# ds_theta_intersect_f() returns an error if it receives an invalid serialized sketch.
|
|
select ds_theta_intersect_f(sketch1, sketch2) from (
|
|
select ds_theta_sketch(float_col) sketch1, max(date_string_col) sketch2 from
|
|
functional_parquet.alltypestiny
|
|
) t
|
|
---- CATCH
|
|
UDF ERROR: Unable to deserialize sketch: sketch type mismatch: expected 3, actual 47
|
|
====
|
|
---- QUERY
|
|
# Intersect the sketches from theta_sketches_impala_hive2 and checks if the intersect
|
|
# produces the same result as if these sketches were used separately to get the estimates.
|
|
select
|
|
ds_theta_estimate(ds_theta_intersect_f(i_ti, h_ti)) as ti,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_i, h_i)) as i,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_bi, h_bi)) as bi,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_f, h_f)) as f,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_d, h_d)) as d,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_s, h_s)) as s,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_c, h_c)) as c,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_v, h_v)) as v,
|
|
ds_theta_estimate(ds_theta_intersect_f(i_nc, h_nc)) as nc
|
|
from theta_sketches_impala_hive2;
|
|
---- TYPES
|
|
BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT,BIGINT
|
|
---- RESULTS
|
|
5,7,6,6,7,4,4,3,0
|
|
====
|
|
---- QUERY
|
|
# Check that the non-empty input sketches are distinct and the result is empty.
|
|
select ds_theta_estimate(ds_theta_intersect_f(date_sketch, float_sketch))
|
|
from sketch_store where year=2009 and month=1;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
0
|
|
====
|
|
---- QUERY
|
|
# Check When the inputs aren't the same but has some (but not all) items common.
|
|
select ds_theta_estimate(ds_theta_intersect_f(sketch1, sketch2)) from sketch_intermediate;
|
|
---- TYPES
|
|
BIGINT
|
|
---- RESULTS
|
|
5
|
|
==== |