Since commit d2d3f4c (on asf-master), TAggregateExpr contains
the logical input types of the Aggregate Expr. The reason they
are included is that merging aggregate expressions will have
input tyes of the intermediate values which aren't necessarily
the same as the input types. For instance, NDV() uses a binary
blob as its intermediate value and it's passed to its merge
aggregate expressions as a StringVal but the input type of NDV()
in the query could be DecimalVal. In this case, we consider
DecimalVal as the logical input type while StringVal is the
intermediate type. The logical input types are accessed by the
BE via GetConstFnAttr() during interpretation and constant
propagation during codegen.
To handle distinct aggregate expressions (e.g. select count(distinct)),
the FE uses 2-phase aggregation by introducing an extra phase of
split/merge aggregation in which the distinct aggregate expressions'
inputs are coverted and added to the group-by expressions in the first
phase while the non-distinct aggregate expressions go through the normal
split/merge treatement.
The bug is that the existing code incorrectly propagates the intermediate
types of the non-grouping aggregate expressions as the logical input types
to the merging aggregate expressions in the second phase of aggregation.
The input aggregate expressions for the non-distinct aggregate expressions
in the second phase aggregation are already merging aggregate expressions
(from phase one) in which case we should not treat its input types as
logical input types.
This change fixes the problem above by checking if the input aggregate
expression passed to FunctionCallExpr.createMergeAggCall() is already
a merging aggregate expression. If so, it will use the logical input
types recorded in its 'mergeAggInputFn_' as references for its logical
input types instead of the aggregate expression input types themselves.
Change-Id: I158303b20d1afdff23c67f3338b9c4af2ad80691
Reviewed-on: http://gerrit.cloudera.org:8080/6724
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Impala Public Jenkins
This uses the existing infrastructure for codegening builtin UDAs and
for codegening calls to UDFs. GetUdf() is refactored to support both
UDFs and UDAs.
IR UDAs are still not allowed by the frontend. It's unclear if we want
to enable them going forward because of the difficulties in testing and
supporting IR UDFs/UDAs.
This also fixes some bugs with the Get*Type() methods of
FunctionContext. GetArgType() and related methods now always return the
logical input types of the aggregate function. Getting the tests to pass
required fixing IMPALA-4878 because they called GetIntermediateType().
Testing:
test_udfs.py tests UDAs with codegen enabled and disabled.
Added some asserts to test UDAs to check that the correct types are
passed in via the FunctionContext.
Change-Id: Id1708eaa96eb76fb9bec5eeabf209f81c88eec2f
Reviewed-on: http://gerrit.cloudera.org:8080/5161
Reviewed-by: Dan Hecht <dhecht@cloudera.com>
Tested-by: Impala Public Jenkins
This change enables codegen for all builtin aggregate functions,
e.g. timestamp functions and group_concat.
There are several parts to the change:
* Adding support for generic UDAs. Previous the codegen code did not
handle multiple input arguments or NULL return values.
* Defaulting to using the UDA interface when there is not a special
codegen path (we have implementations of all builtin aggregate
functions for the interpreted path).
* Remove all the logic to disable codegen for the special cases that now
are supported.
Also fix the generation of code to get/set NULL bits since I needed
to add functionality there anyway.
Testing:
Add tests that check that codegen was enabled for builtin aggregate
functions. Also fix some gaps in the preexisting tests.
Also add tests for UDAs that check input/output nulls are handled
correctly, in anticipation of enabling codegen for arbitrary UDAs.
The tests are run with both codegen enabled and disabled. To avoid
flaky tests, we switch the UDF tests to use "unique_database".
Perf:
Ran local TPC-H and targeted perf. Spent a lot of time on TPC-H Q1,
since my original approach regressed it ~5%. In the end the problem was
to do with the ordering of loads/stores to the slot and null bit in the
generated code: the previous version of the code exploited some
properties of the particular aggregate function. I ended up replicating
this behaviour to avoid regressing perf.
Change-Id: Id9dc21d1d676505d3617e1e4f37557397c4fb260
Reviewed-on: http://gerrit.cloudera.org:8080/4655
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Internal Jenkins
This patch fixes two issues around handling of constant expr args.
The patches are combined because they touch some of the same code
and depend on some of the same memory management cleanup.
First, it fixes IMPALA-2379, where constant expr args were not visible
to UDAFs. The issue is that the input exprs need to be opened before
calling the UDAF Init() function.
Second, it avoids overhead from repeated evaluation of constant
arguments for ScalarFnCall expressions on both the codegen'd and
interpreted paths. A common example is an IN predicate with a
long list of constant values.
The interpreted path was inefficient because it always evaluated all
children expressions. Instead in this patch constant args are
evaluated once and cached. The memory management of the AnyVal*
objects was somewhat nebulous - adjusted it so that they're allocated
from ExprContext::mem_pool_, which has the correct lifetime.
The codegen'd path was inefficient only with varargs - with fixed
arguments the LLVM optimiser is able to infer after inlining that
the expressions are constant and remove all evaluation. However,
for varargs it stores the vararg values into a heap-allocated buffer.
The LLVM optimiser is unable to remove these stores because they
have a side-effect that is visible to code outside the function.
The codegen'd path is improved by evaluating varargs into an automatic
buffer that can be optimised out. We also make a small related change
to bake the string constants into the codegen'd code.
Testing:
Ran exhaustive build.
Added regression test for IMPALA-2379 and MemPool test for aligned
allocation. Added a test for in predicates with constant strings.
Perf:
Added a targeted query that demonstrates the improvement. Also manually
validated the non-codegend perf. Also ran TPC-H and targeted perf
queries locally - didn't see any significant changes.
+--------------------+-------------------------------+-----------------------+--------+-------------+------------+-----------+----------------+-------------+-------+
| Workload | Query | File Format | Avg(s) | Base Avg(s) | Delta(Avg) | StdDev(%) | Base StdDev(%) | Num Clients | Iters |
+--------------------+-------------------------------+-----------------------+--------+-------------+------------+-----------+----------------+-------------+-------+
| TARGETED-PERF(_20) | primitive_filter_in_predicate | parquet / none / none | 1.19 | 9.82 | I -87.85% | 3.82% | 0.71% | 1 | 10 |
+--------------------+-------------------------------+-----------------------+--------+-------------+------------+-----------+----------------+-------------+-------+
(I) Improvement: TARGETED-PERF(_20) primitive_filter_in_predicate [parquet / none / none] (9.82s -> 1.19s [-87.85%])
+--------------+------------+----------+----------+------------+-----------+----------+----------+------------+--------+--------+-----------+
| Operator | % of Query | Avg | Base Avg | Delta(Avg) | StdDev(%) | Max | Base Max | Delta(Max) | #Hosts | #Rows | Est #Rows |
+--------------+------------+----------+----------+------------+-----------+----------+----------+------------+--------+--------+-----------+
| 01:AGGREGATE | 14.39% | 155.88ms | 214.61ms | -27.37% | 2.68% | 163.38ms | 227.53ms | -28.19% | 1 | 1 | 1 |
| 00:SCAN HDFS | 85.60% | 927.46ms | 9.43s | -90.16% | 4.49% | 1.01s | 9.50s | -89.42% | 1 | 13.77K | 14.05K |
+--------------+------------+----------+----------+------------+-----------+----------+----------+------------+--------+--------+-----------+
Change-Id: I45c3ed8c9d7a61e94a9b9d6c316e8a53d9ff6c24
Reviewed-on: http://gerrit.cloudera.org:8080/4838
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
Tested-by: Internal Jenkins
Previously the frontend rejected UDAs with different intermediate and
result type. The backend supports these, so this change enables support
in the frontend and adds tests.
This patch adds a test UDA function with different intermediate type and
a simple end-to-end test that exercises it. It modifies an existing
unused test UDA that used a currently unsupported intermediate type -
BufferVal.
Change-Id: I5675ec7f275ea698c24ea8e92de7f469a950df83
Reviewed-on: http://gerrit.cloudera.org:8080/655
Tested-by: Internal Jenkins
Reviewed-by: Tim Armstrong <tarmstrong@cloudera.com>
The following changes are included in this commit:
1. Modified the alltypesagg table to include an additional partition key
that has nulls.
2. Added a number of tests in hdfs.test that exercise the partition
pruning logic (see IMPALA-887).
3. Modified all the tests that are affected by the change in alltypesagg.
Change-Id: I1a769375aaa71273341522eb94490ba5e4c6f00d
Reviewed-on: http://gerrit.ent.cloudera.com:8080/2874
Reviewed-by: Dimitris Tsirogiannis <dtsirogiannis@cloudera.com>
Tested-by: jenkins
Reviewed-on: http://gerrit.ent.cloudera.com:8080/3236
