On Thu, Aug 10, 2017 at 6:23 PM, Ashutosh Bapat
wrote:
> Your patch didn't improve planning time without partition-wise join,
> so it's something good to have along-with partition-wise join. Given
> that Bitmapsets are used in other parts of code as well, the
>
On Thu, Aug 10, 2017 at 9:28 AM, Thomas Munro
wrote:
> On Thu, Aug 10, 2017 at 1:39 AM, Thomas Munro
> wrote:
>> On my computer it took ~1.5 seconds to plan a 1000 partition join,
>> ~7.1 seconds to plan a 2000 partition join, and ~50
On Thu, Aug 10, 2017 at 1:39 AM, Thomas Munro
wrote:
> On my computer it took ~1.5 seconds to plan a 1000 partition join,
> ~7.1 seconds to plan a 2000 partition join, and ~50 seconds to plan a
> 4000 partition join. I poked around in a profiler a bit and saw that
On Tue, Aug 8, 2017 at 8:51 PM, Ashutosh Bapat
wrote:
> Updated patches attached.
Hi,
I started reviewing this. It's nicely commented, but it's also very
complicated, and it's going to take me several rounds to understand
what all the parts do, but here's some
On Thu, Aug 3, 2017 at 9:38 PM, Ashutosh Bapat
wrote:
> On Thu, Aug 3, 2017 at 2:10 AM, Robert Haas wrote:
>> On Mon, Jul 31, 2017 at 7:59 AM, Ashutosh Bapat
>> wrote:
>>> Adding AppendRelInfos to
On Thu, Aug 3, 2017 at 2:10 AM, Robert Haas wrote:
> On Mon, Jul 31, 2017 at 7:59 AM, Ashutosh Bapat
> wrote:
>> Adding AppendRelInfos to root->append_rel_list as and when they are
>> created would keep parent AppendRelInfos before those of
On Mon, Jul 31, 2017 at 9:07 AM, Ashutosh Bapat
wrote:
> Forgot the patch set. Here it is.
The commit message for 0005 isn't really accurate given that it
follows 0004. I think you could just flatten 0005 and 0006 into one
patch.
Reviewing those together:
-
On Mon, Jul 31, 2017 at 7:59 AM, Ashutosh Bapat
wrote:
> Adding AppendRelInfos to root->append_rel_list as and when they are
> created would keep parent AppendRelInfos before those of children. But
> that function throws away the AppendRelInfo it created when
On Mon, Jul 31, 2017 at 8:32 AM, Robert Haas wrote:
> On Fri, Jul 14, 2017 at 3:02 AM, Ashutosh Bapat
> wrote:
>> Here's revised patch set with only 0004 revised. That patch deals with
>> creating multi-level inheritance hierarchy from
On Fri, Jul 14, 2017 at 3:02 AM, Ashutosh Bapat
wrote:
> Here's revised patch set with only 0004 revised. That patch deals with
> creating multi-level inheritance hierarchy from multi-level partition
> hierarchy. The original logic of recursively calling
>
On Wed, Jul 26, 2017 at 12:02 PM, Ashutosh Bapat
wrote:
>
> Ok. If those queries have equi-join between partitioned tables and are
> not picking up partition-wise join, that case needs to be
> investigated. Q21 for example has join between three lineitem
>
On Wed, Jul 26, 2017 at 10:38 AM, Ashutosh Bapat
wrote:
>
> On Tue, Jul 25, 2017 at 9:39 PM, Dilip Kumar wrote:
> > On Tue, Jul 25, 2017 at 8:59 PM, Robert Haas wrote:
> >> On Tue, Jul 25, 2017 at 1:31 AM, Rafia
On Wed, Jul 26, 2017 at 11:08 AM, Rafia Sabih
wrote:
>
>
> On Wed, Jul 26, 2017 at 11:06 AM, Ashutosh Bapat
> wrote:
>>
>> On Wed, Jul 26, 2017 at 11:00 AM, Rafia Sabih
>> wrote:
>> >
>> >
>> > On Wed,
On Wed, Jul 26, 2017 at 11:06 AM, Ashutosh Bapat <
ashutosh.ba...@enterprisedb.com> wrote:
> On Wed, Jul 26, 2017 at 11:00 AM, Rafia Sabih
> wrote:
> >
> >
> > On Wed, Jul 26, 2017 at 10:58 AM, Ashutosh Bapat
> > wrote:
> >>
> >> On
On Wed, Jul 26, 2017 at 11:00 AM, Rafia Sabih
wrote:
>
>
> On Wed, Jul 26, 2017 at 10:58 AM, Ashutosh Bapat
> wrote:
>>
>> On Tue, Jul 25, 2017 at 11:01 AM, Rafia Sabih
>> wrote:
>>
>> > Query plans for
On Wed, Jul 26, 2017 at 10:58 AM, Ashutosh Bapat <
ashutosh.ba...@enterprisedb.com> wrote:
> On Tue, Jul 25, 2017 at 11:01 AM, Rafia Sabih
> wrote:
>
> > Query plans for the above mentioned queries is attached.
> >
>
> Can you please share plans for all the queries,
On Tue, Jul 25, 2017 at 11:01 AM, Rafia Sabih
wrote:
> Query plans for the above mentioned queries is attached.
>
Can you please share plans for all the queries, even if they haven't
chosen partition-wise join when run on partitioned tables with
On Tue, Jul 25, 2017 at 9:39 PM, Dilip Kumar wrote:
> On Tue, Jul 25, 2017 at 8:59 PM, Robert Haas wrote:
>> On Tue, Jul 25, 2017 at 1:31 AM, Rafia Sabih
>> wrote:
>>> - other queries show a good 20-30% improvement in
On Tue, Jul 25, 2017 at 8:59 PM, Robert Haas wrote:
> On Tue, Jul 25, 2017 at 1:31 AM, Rafia Sabih
> wrote:
>> - other queries show a good 20-30% improvement in performance. Performance
>> numbers are as follows,
>>
>> Query| un_part_head
On Tue, Jul 25, 2017 at 1:31 AM, Rafia Sabih
wrote:
> - other queries show a good 20-30% improvement in performance. Performance
> numbers are as follows,
>
> Query| un_part_head (seconds) | part_head (seconds) | part_patch (seconds) |
> 3 | 76 |127 | 88 |
> 4 |17 |
On Fri, Jul 21, 2017 at 11:54 AM, Rafia Sabih
wrote:
> So, does this
> also mean that a partitioned table will not join with an unpartitioned
> table without append of partitions?
>
Yes. When you join an unpartitioned table with a partitioned table,
the planner will
On Thu, Jul 20, 2017 at 2:44 PM, Ashutosh Bapat
wrote:
>
> On Thu, Jul 20, 2017 at 11:46 AM, Amit Langote
> wrote:
> > On 2017/07/20 15:05, Ashutosh Bapat wrote:
> >> On Wed, Jul 19, 2017 at 9:54 AM, Rafia Sabih
> >>
On Fri, Jul 21, 2017 at 11:42 AM, Rafia Sabih
wrote:
>
> Following the discussion at [1], with the patch Thomas posted there, now Q21
> completes in some 160 seconds.
Your earlier reports mentioned unpartitioned case taking 300 seconds,
partitioned case without
On Thu, Jul 20, 2017 at 8:53 AM, Thomas Munro wrote:
> On Thu, Jul 20, 2017 at 2:02 PM, Robert Haas
> wrote:
> > On Wed, Jul 19, 2017 at 7:45 PM, Thomas Munro
> > wrote:
> >> Isn't this the same as the issue
On Thu, Jul 20, 2017 at 11:46 AM, Amit Langote
wrote:
> On 2017/07/20 15:05, Ashutosh Bapat wrote:
>> On Wed, Jul 19, 2017 at 9:54 AM, Rafia Sabih
>> wrote:
>>>
>>> Partition information:
>>> Type of partitioning - single column range
On Thu, Jul 20, 2017 at 12:30 AM, Robert Haas wrote:
>
>> This suggests that partitioning is not a suitable strategy for this query,
>> but then may be partition wise should not be picked for such a case to
>> aggravate the performance issue.
>
> In the unpartitioned case,
On 2017/07/20 15:05, Ashutosh Bapat wrote:
> On Wed, Jul 19, 2017 at 9:54 AM, Rafia Sabih
> wrote:
>>
>> Partition information:
>> Type of partitioning - single column range partition
>> Tables partitioned - Lineitem and orders
>>
>> Lineitem -
>> Partition key =
On Wed, Jul 19, 2017 at 9:54 AM, Rafia Sabih
wrote:
>
> Partition information:
> Type of partitioning - single column range partition
> Tables partitioned - Lineitem and orders
>
> Lineitem -
> Partition key = l_orderkey
> No of partitions = 18
>
> Orders -
>
On Thu, Jul 20, 2017 at 2:02 PM, Robert Haas wrote:
> On Wed, Jul 19, 2017 at 7:45 PM, Thomas Munro
> wrote:
>> Isn't this the same as the issue reported here?
>>
>>
On Wed, Jul 19, 2017 at 7:45 PM, Thomas Munro
wrote:
> On Thu, Jul 20, 2017 at 7:00 AM, Robert Haas wrote:
>> I think the problem is that the row count estimates for the child
>> joins seem to be totally bogus:
>>
>> -> Hash Semi Join
On Thu, Jul 20, 2017 at 7:00 AM, Robert Haas wrote:
> I think the problem is that the row count estimates for the child
> joins seem to be totally bogus:
>
> -> Hash Semi Join (cost=309300.53..491665.60 rows=1 width=12)
> (actual time=10484.422..15945.851 rows=1523493
On Wed, Jul 19, 2017 at 12:24 AM, Rafia Sabih
wrote:
> On testing this patch for TPC-H (for scale factor 20) benchmark I found a
> regression for Q21, on head it was taking some 600 seconds and with this
> patch it is taking 3200 seconds. This comparison is on the
On Fri, Jun 30, 2017 at 2:53 PM, Rafia Sabih
wrote:
>
>
> On Mon, May 22, 2017 at 12:02 PM, Ashutosh Bapat
> wrote:
>>
>>
>> Here's set of patches rebased on latest head.
>
>
> In an attempt to test this set of patches, I found that
On Mon, May 22, 2017 at 12:02 PM, Ashutosh Bapat <
ashutosh.ba...@enterprisedb.com> wrote:
>
> Here's set of patches rebased on latest head.
>
In an attempt to test this set of patches, I found that not all of the
patches could be applied on latest head-- commit
On Sat, Apr 29, 2017 at 12:37 AM, Robert Haas wrote:
> On Fri, Apr 28, 2017 at 1:18 AM, Ashutosh Bapat
> wrote:
>> For two-way join this works and is fairly straight-forward. I am
>> assuming that A an B are base relations and not joins.
On Thu, May 18, 2017 at 4:38 AM, Ashutosh Bapat
wrote:
> But in case a user has written an = operator which returns true for
> two NULL values, per description in [1], that comparison operator is
> flawed and
> using that operator is going to result in
On Thu, Apr 6, 2017 at 6:37 AM, Robert Haas wrote:
>
>> There's a relevant comment in 0006, build_joinrel_partition_info()
>> (probably that name needs to change, but I will do that once we have
>> settled on design)
>> + /*
>> +* Construct partition keys for the
On Fri, Apr 28, 2017 at 1:18 AM, Ashutosh Bapat
wrote:
> For two-way join this works and is fairly straight-forward. I am
> assuming that A an B are base relations and not joins. But making it
> work for N-way join is the challenge.
I don't think it's much
On Fri, Apr 28, 2017 at 1:32 AM, Robert Haas wrote:
> On Thu, Apr 27, 2017 at 3:41 AM, Ashutosh Bapat
> wrote:
>> The third goal requires that the partition bounds be compared based on
>> the partition keys present in the equi-join. While
On Thu, Apr 27, 2017 at 3:41 AM, Ashutosh Bapat
wrote:
> The third goal requires that the partition bounds be compared based on
> the partition keys present in the equi-join. While matching the
> partitions to be joined, the partition bounds corresponding the base
On Wed, Apr 26, 2017 at 9:30 PM, Robert Haas wrote:
> On Mon, Apr 24, 2017 at 7:06 AM, Ashutosh Bapat
> wrote:
>> This assumes that datums in partition bounds have one to one mapping
>> with the partitions, which isn't true for list
On Wed, Apr 26, 2017 at 12:41 PM, Tom Lane wrote:
> Robert Haas writes:
>> So this is about a cross-type join,
>> not multiple types within a single partitioning hierarchy, as you
>> might also gather from the subject line of this thread.
>
> OK, but I
Robert Haas writes:
> So this is about a cross-type join,
> not multiple types within a single partitioning hierarchy, as you
> might also gather from the subject line of this thread.
OK, but I still don't understand why any type conversion is needed
in such a case. The
On Wed, Apr 26, 2017 at 12:19 PM, Tom Lane wrote:
> What I'm going to ask one more time, though, is why we are even discussing
> this. Surely the partition bounds of a partitioned table must all be of
> the same type already. If there is a case where they are not, that is
>
Robert Haas writes:
> I'm going to say this one more time: I really, really, really think
> you need to avoid trying to convert the partition bounds to a common
> type. I said before that the infrastructure to do that is not present
> in our type system, and I'm pretty
On Mon, Apr 24, 2017 at 7:06 AM, Ashutosh Bapat
wrote:
> This assumes that datums in partition bounds have one to one mapping
> with the partitions, which isn't true for list partitions. For list
> partitions we have multiple datums corresponding to the items
On Mon, Apr 24, 2017 at 5:02 PM, Rajkumar Raghuwanshi
wrote:
> On Fri, Apr 21, 2017 at 7:59 PM, Ashutosh Bapat
> wrote:
>>
>> Here's an updated patch set
>
>
> Hi,
>
> I have applied v18 patches and got a crash in m-way
On Fri, Apr 21, 2017 at 7:59 PM, Ashutosh Bapat <
ashutosh.ba...@enterprisedb.com> wrote:
> Here's an updated patch set
>
Hi,
I have applied v18 patches and got a crash in m-way joins when partition
ranges differ, below are steps to reproduce this.
CREATE TABLE prt1 (a int, b int, c varchar)
On Sat, Apr 22, 2017 at 3:52 AM, Robert Haas wrote:
> On Fri, Apr 21, 2017 at 8:41 AM, Ashutosh Bapat
> wrote:
>> I don't see how is that fixed. For a join relation we need to come up
>> with one set of partition bounds by merging partition
On Fri, Apr 21, 2017 at 8:41 AM, Ashutosh Bapat
wrote:
> I don't see how is that fixed. For a join relation we need to come up
> with one set of partition bounds by merging partition bounds of the
> joining relation and in order to understand how to interpret the
On Fri, Apr 21, 2017 at 1:34 AM, Robert Haas wrote:
>> You seem to be suggesting that we keep as many sets of
>> partition bounds as there are base relations participating in the join
>> and then use appropriate partition bounds based on the columns in the
>> join
On Thu, Apr 20, 2017 at 8:45 AM, Ashutosh Bapat
wrote:
> I think you are confusing join condition application and partition
> bounds of a join relation.
You're right, I misunderstood what you were talking about.
> But the problem we are trying to solve here
On Thu, Apr 20, 2017 at 3:35 PM, Amit Langote
wrote:
> On 2017/04/20 15:45, Ashutosh Bapat wrote:
>> On Thu, Apr 20, 2017 at 10:42 AM, Robert Haas wrote:
>>> I don't understand why you think that partition-wise join needs any
>>> new logic
On 2017/04/20 15:45, Ashutosh Bapat wrote:
> On Thu, Apr 20, 2017 at 10:42 AM, Robert Haas wrote:
>> I don't understand why you think that partition-wise join needs any
>> new logic here; if this were a non-partitionwise join, we'd similarly
>> need to use the correct
On Thu, Apr 20, 2017 at 10:42 AM, Robert Haas wrote:
> On Tue, Apr 18, 2017 at 6:55 AM, Ashutosh Bapat
> wrote:
>> When we merge partition bounds from two relations with different
>> partition key types, the merged partition bounds need to
On Thu, Apr 20, 2017 at 11:32 AM, Tom Lane wrote:
>
> BTW, I remain totally mystified as to what people think the semantics of
> partitioning ought to be. Child columns can have a different type from
> parent columns? Really? Why is this even under discussion? We don't
>
Robert Haas writes:
> I don't understand why you think that partition-wise join needs any
> new logic here; if this were a non-partitionwise join, we'd similarly
> need to use the correct operator, but the existing code handles that
> just fine. If the join is performed
On Tue, Apr 18, 2017 at 6:55 AM, Ashutosh Bapat
wrote:
> When we merge partition bounds from two relations with different
> partition key types, the merged partition bounds need to have some
> information abound the way those constants look like e.g. their
>
On Wed, Apr 5, 2017 at 8:39 AM, Robert Haas wrote:
> On Tue, Apr 4, 2017 at 10:22 AM, Ashutosh Bapat
> wrote:
>> Yes, I agree. For an inner join, the partition key types need to "shrink"
>> and for outer join they need to be "widened". I
On Thu, Apr 6, 2017 at 6:37 AM, Robert Haas wrote:
> On Wed, Apr 5, 2017 at 2:42 AM, Ashutosh Bapat
> wrote:
>> Only inner join conditions have equivalence classes associated with
>> those. Outer join conditions create single element
Robert Haas writes:
> ... But the underlying point here is that
> the only thing you really know about the function is that it's got to
> be a strategy-3 operator in some btree opclass; if that guarantees
> strictness, then so be it -- but I wasn't able to find anything in
On Wed, Apr 5, 2017 at 2:42 AM, Ashutosh Bapat
wrote:
> Only inner join conditions have equivalence classes associated with
> those. Outer join conditions create single element equivalence
> classes. So, we can not associate equivalence classes as they are with
>
On Wed, Apr 5, 2017 at 8:39 AM, Robert Haas wrote:
> On Tue, Apr 4, 2017 at 10:22 AM, Ashutosh Bapat
> wrote:
>> Yes, I agree. For an inner join, the partition key types need to "shrink"
>> and for outer join they need to be "widened". I
On Tue, Apr 4, 2017 at 10:22 AM, Ashutosh Bapat
wrote:
> Yes, I agree. For an inner join, the partition key types need to "shrink"
> and for outer join they need to be "widened". I don't know if there is a way
> to know "wider" or "shorter" of two given types. We
On Thu, Mar 30, 2017 at 1:14 AM, Ashutosh Bapat
wrote:
> Done.
Ashutosh and I spent several hours discussing this patch set today.
I'm starting to become concerned about the fact that 0004 makes the
partition bounds part of the PartitionScheme, because that means
On Thu, Mar 30, 2017 at 6:32 AM, Amit Langote
wrote:
> So, because we want to create an Append path for each partitioned table in
> a tree separately, we'll need RelOptInfo for each one, which in turn
> requires an AppendRelInfo.
Hmm. It's no more desirable to
On 2017/03/30 18:35, Ashutosh Bapat wrote:
>> On Wed, Mar 29, 2017 at 8:39 AM, Robert Haas wrote:
>> I don't think 0011 is likely to be acceptable in current form. I
>> can't imagine that we just went to the trouble of getting rid of
>> AppendRelInfos for child partitioned
> On Wed, Mar 29, 2017 at 8:39 AM, Robert Haas wrote:
> I don't think 0011 is likely to be acceptable in current form. I
> can't imagine that we just went to the trouble of getting rid of
> AppendRelInfos for child partitioned rels only to turn around and put
> them back
On Tue, Mar 28, 2017 at 10:24 PM, Robert Haas wrote:
> On Mon, Mar 27, 2017 at 8:36 AM, Ashutosh Bapat
> wrote:
>> I have gone through the patch, and it looks good to me. Here's the set
>> of patches with this patch included. Fixed the
On Tue, Mar 28, 2017 at 12:54 PM, Robert Haas wrote:
> Regarding 0002, I think the parts that involve factoring out
> find_param_path_info() are uncontroversial. Regarding the changes to
> adjust_appendrel_attrs(), my main question is whether we wouldn't be
> better off
On Mon, Mar 27, 2017 at 8:36 AM, Ashutosh Bapat
wrote:
> I have gone through the patch, and it looks good to me. Here's the set
> of patches with this patch included. Fixed the testcase failures.
> Rebased the patchset on de4da168d57de812bb30d359394b7913635d21a9.
Hi Ashutosh,
On 2017/03/23 21:48, Ashutosh Bapat wrote:
>>> I have fixed all the issues reported till now.
I've tried to fix your 0012 patch (Multi-level partitioned table
expansion) considering your message earlier on this thread [1].
Especially the fact that no AppendRelInfo and RelOptInfo are
Hi Ashutosh,
On 2017/03/23 21:48, Ashutosh Bapat wrote:
>>> I have fixed all the issues reported till now.
In patch 0007, the following code in have_partkey_equi_join() looks
potentially unsafe:
/*
* If the clause refers to different partition keys from
* both
>
>> I don't think we will get away by supporting just scan paths, since
>> the inner side of lateral join can be any paths not just scan path. Or
>> you are suggesting that we disable partition-wise lateral join and
>> support reparameterization of only scan paths?
>
> I think if you can do a
On Wed, Mar 22, 2017 at 8:46 AM, Ashutosh Bapat
wrote:
> Here's set of updated patches rebased on
> 1148e22a82edc96172fc78855da392b6f0015c88.
>
> I have fixed all the issues reported till now.
I don't understand why patch 0001 ends up changing every existing test
On Wed, Mar 22, 2017 at 9:59 AM, Ashutosh Bapat
wrote:
>> In my testing last week, reparameterize_path_by_child() was essential
>> for nested loops to work properly, even without LATERAL. Without it,
>> the parameterized path ends up containing vars that
On Wed, Mar 22, 2017 at 6:32 PM, Robert Haas wrote:
> On Wed, Mar 22, 2017 at 8:46 AM, Ashutosh Bapat
> wrote:
>> I have also completed reparameterize_path_by_child() for all the
>> required paths. There's no TODO there now. :) The function
On Wed, Mar 22, 2017 at 8:46 AM, Ashutosh Bapat
wrote:
> I have also completed reparameterize_path_by_child() for all the
> required paths. There's no TODO there now. :) The function has grown
> quite long now and might take some time to review. Given the size, I
>
> I tested the applied patch, it is fixing the reported issue.
Thanks for the confirmation Rafia. I have included the fix in the
latest set of patches.
--
Best Wishes,
Ashutosh Bapat
EnterpriseDB Corporation
The Postgres Database Company
--
Sent via pgsql-hackers mailing list
On Wed, Mar 22, 2017 at 3:19 PM, Ashutosh Bapat
wrote:
>> In an attempt to test the geqo side of this patch, I reduced
>> geqo_threshold to 6 and set enable_partitionwise_join to to true and
>> tried following query, which crashed,
>>
>> explain select * from
>>>
> In an attempt to test the geqo side of this patch, I reduced
> geqo_threshold to 6 and set enable_partitionwise_join to to true and
> tried following query, which crashed,
>
> explain select * from prt, prt2, prt3, prt32, prt4, prt42 where prt.a
> = prt2.b and prt3.a = prt32.b and prt4.a =
On Tue, Mar 21, 2017 at 10:40 PM, Ashutosh Bapat
wrote:
> Here's the set of patches rebased on latest head, which also has a
> commit to eliminate scans on partitioned tables. This change has
> caused problems with multi-level partitioned tables, that I have not
>
On Tue, Mar 21, 2017 at 7:41 AM, Ashutosh Bapat
wrote:
> On Mon, Mar 20, 2017 at 10:17 PM, Ashutosh Bapat
> wrote:
>>>
>>> On a further testing of this patch I find another case when it is
>>> showing regression, the time taken
Thanks Rajkumar. Added those in the latest set of patches.
On Tue, Mar 21, 2017 at 3:52 PM, Rajkumar Raghuwanshi
wrote:
>> On Mon, Mar 20, 2017 at 1:19 PM, Ashutosh Bapat
>> wrote:
>
> I have created some test to cover
> On Mon, Mar 20, 2017 at 1:19 PM, Ashutosh Bapat
> wrote:
I have created some test to cover partition wise joins with
postgres_fdw, also verified make check.
patch attached.
Thanks & Regards,
Rajkumar Raghuwanshi
QMG, EnterpriseDB Corporation
diff --git
On Mon, Mar 20, 2017 at 1:19 PM, Ashutosh Bapat
wrote:
>> That seems different than what I suggested and I'm not sure what the
>> reason is for the difference?
>
> The patch adding macros IS_JOIN_REL() and IS_OTHER_REL() and changing
> the code to use it will look
On Mon, Mar 20, 2017 at 10:26 PM, Robert Haas wrote:
> On Mon, Mar 20, 2017 at 12:52 PM, Ashutosh Bapat
> wrote:
>>> Hmm. I would kind of like to move the IS_JOIN_REL() and
>>> IS_OTHER_REL() stuff to the front of the series. In other
On Mon, Mar 20, 2017 at 12:52 PM, Ashutosh Bapat
wrote:
>> Hmm. I would kind of like to move the IS_JOIN_REL() and
>> IS_OTHER_REL() stuff to the front of the series. In other words, I
>> propose that we add those macros first, each testing for only the one
>>
>
> On a further testing of this patch I find another case when it is
> showing regression, the time taken with patch is around 160 secs and
> without it is 125 secs.
> Another minor thing to note that is planning time is almost twice with
> this patch, though I understand that this is for
>
> Hmm. I would kind of like to move the IS_JOIN_REL() and
> IS_OTHER_REL() stuff to the front of the series. In other words, I
> propose that we add those macros first, each testing for only the one
> kind of RelOptInfo that exists today, and change all the code to use
> them. Then, when we
On Mon, Mar 20, 2017 at 12:07 PM, Rafia Sabih
wrote:
> On a further testing of this patch I find another case when it is
> showing regression, the time taken with patch is around 160 secs and
> without it is 125 secs.
This is basically the same problem as before;
On Mon, Mar 20, 2017 at 9:44 AM, Ashutosh Bapat
wrote:
> Right. If we could use parent Vars to indicate parent Var or child Var
> depending upon the context, a lot of memory issues would be solved; we
> wouldn't need to translate a single expression. But I think
On Mon, Mar 20, 2017 at 9:44 AM, Ashutosh Bapat
wrote:
>> I believe it would also be best to include 0011's changes to
>> adjust_appendrel_attrs_multilevel in 0001.
>
> The function needs to repeat the "adjustment" process for every
> "other" relation (join or
On Mon, Mar 20, 2017 at 8:21 AM, Robert Haas wrote:
> On Fri, Mar 17, 2017 at 8:10 PM, Robert Haas wrote:
>> While I was studying what you did with reparameterize_path_by_child(),
>> I started to wonder whether reparameterize_path() doesn't need to
On Sat, Mar 18, 2017 at 5:40 AM, Robert Haas wrote:
> On Fri, Mar 17, 2017 at 9:15 AM, Ashutosh Bapat
> wrote:
>> This set of patches fixes both of those things.
>
> 0001 changes the purpose of a function and then 0007 renames it. It
>
On Fri, Mar 17, 2017 at 8:10 PM, Robert Haas wrote:
> While I was studying what you did with reparameterize_path_by_child(),
> I started to wonder whether reparameterize_path() doesn't need to
> start handling join paths. I think it only handles scan paths right
> now
On Sun, Mar 19, 2017 at 12:15 AM, Rafia Sabih
wrote:
> I was trying to play around with this patch and came across following
> case when without the patch query completes in 9 secs and with it in
> 15 secs. Theoretically, I tried to capture the case when each
>
On Sat, Mar 18, 2017 at 5:40 AM, Robert Haas wrote:
> On Fri, Mar 17, 2017 at 9:15 AM, Ashutosh Bapat
> wrote:
>> This set of patches fixes both of those things.
>
> 0001 changes the purpose of a function and then 0007 renames it. It
>
On Fri, Mar 17, 2017 at 9:15 AM, Ashutosh Bapat
wrote:
> This set of patches fixes both of those things.
0001 changes the purpose of a function and then 0007 renames it. It
would be better to include the renaming in 0001 so that you're not
taking multiple whacks
On Thu, Mar 16, 2017 at 7:19 AM, Ashutosh Bapat
wrote:
> On Thu, Mar 16, 2017 at 7:10 AM, Robert Haas wrote:
>> So I am looking at this part of 0008:
>>
>> + /*
>> +* Do not copy parent_rinfo and child_rinfos because 1. they
101 - 200 of 240 matches
Mail list logo