Re: [HACKERS] path toward faster partition pruning

On 6 April 2018 at 10:35, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> I changed a lot of code also, but cosmetic changes only.
>
> I'll clean this up a bit more now, and try to commit shortly (or early
> tomorrow); wanted to share current status now in case I have to rush
> out.

I made a complete pass over the patch you sent. I only noted down the
following few things:

1.

+ * off < 0, meaning the look-up value is smaller that all bounds,

that -> than

2. I guess this will be removed before commit.

+#if 0
<large section of comments>
+#endif

3. This comment seems like a strange thing to write just before
testing if the clause matches the partition key.

+ /* Clause does not match this partition key. */
+ if (equal(leftop, partkey))
+ *rightop = not_clause((Node *) clause)
+ ? (Expr *) makeBoolConst(false, false)
+ : (Expr *) makeBoolConst(true, false);

4. Comment needs removed.

+ * has_default_part - Whether the table has a default partition

The only other thing I noted on this pass is that we could get rid of:

+ /* go check the next clause. */
+ if (unsupported_clause)
+ break;

and just "continue" instead of "break" in all cases apart from case
PARTCLAUSE_UNSUPPORTED:

it would save a few lines and a single condition. What's there works,
but thought this might be better...

--
David Rowley http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services