Re: 回复:Re: Re: 回复:Re: Is it worth pushing conditions to sublink/subplan? - Mailing list pgsql-hackers

From Zhihong Yu
Subject Re: 回复:Re: Re: 回复:Re: Is it worth pushing conditions to sublink/subplan?
Date
Msg-id CALNJ-vTJiOZLEgLG6foxZ8UOPg2yikRD+ovtQiiZQEfXupaqUw@mail.gmail.com
Whole thread Raw
In response to 回复:回复:Re: Re: 回复:Re: Is it worth pushing conditions to sublink/subplan?  ("曾文旌(义从)" <wenjing.zwj@alibaba-inc.com>)
Responses Re: 回复:Re: Re: 回复:Re: Is it worth pushing conditions to sublink/subplan?  (wenjing <wjzeng2012@gmail.com>)
List pgsql-hackers


On Thu, Dec 23, 2021 at 3:52 AM 曾文旌(义从) <wenjing.zwj@alibaba-inc.com> wrote:

Fixed a bug found during testing.


Wenjing

Hi,
+           if (condition_is_safe_pushdown_to_sublink(rinfo, expr_info->outer))
+           {
+               /* replace qual expr from outer var = const to var = const and push down to sublink query */
+               sublink_query_push_qual(subquery, (Node *)copyObject(rinfo->clause), expr_info->outer, expr_info->inner);

Since sublink_query_push_qual() is always guarded by condition_is_safe_pushdown_to_sublink(), it seems sublink_query_push_qual() can be folded into condition_is_safe_pushdown_to_sublink().

For generate_base_implied_equalities():

+       if (ec->ec_processed)
+       {
+           ec_index++;
+           continue;
+       }
+       else if (list_length(ec->ec_members) > 1)

Minor comment: the keyword `else` can be omitted (due to `continue` above).

+            * Since there may be an unexpanded sublink in the targetList,
+            * we'll skip it for now.

Since there may be an -> If there is an

+       {"lazy_process_sublink", PGC_USERSET, QUERY_TUNING_METHOD,
+           gettext_noop("enable lazy process sublink."),

Looking at existing examples from src/backend/utils/misc/guc.c, enable_lazy_sublink_processing seems to be consistent with existing guc variable naming.

+lazy_process_sublinks(PlannerInfo *root, bool single_result_rte)

lazy_process_sublinks -> lazily_process_sublinks

+   else
+   {
    /* There shouldn't be any OJ info to translate, as yet */
    Assert(subroot->join_info_list == NIL);

Indentation for the else block is off.

+       if (istop)
+           f->quals = preprocess_expression_ext(root, f->quals, EXPRKIND_QUAL, false);
+       else
+           f->quals = preprocess_expression_ext(root, f->quals, EXPRKIND_QUAL, true);

The above can be written as:

+           f->quals = preprocess_expression_ext(root, f->quals, EXPRKIND_QUAL, !istop);

For find_equal_conditions_contain_uplevelvar_in_sublink_query():
+               context.has_unexpected_expr == false &&
`!context.has_unexpected_expr` should suffice

equal_expr_safety_check -> is_equal_expr_safe

Cheers

pgsql-hackers by date:

Previous
From: Bharath Rupireddy
Date:
Subject: Re: add recovery, backup, archive, streaming etc. activity messages to server logs along with ps display
Next
From: Andrew Dunstan
Date:
Subject: Re: Proposal: sslmode=tls-only