Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Fabien COELHO 



I don't want to go too deep into it, but you get stuff like this:

Select pow(2.0, -3)::text = pow(2, -3)::text;


Sure. It does so with any overloaded operator or function:

  fabien=# SELECT (2.0 + 3)::TEXT = (2 + 3)::TEXT; # f

Patch applies, make check ok in pgbench, doc gen ok.

ipow code is nice and simple.

I switched the patch to "Ready for Committer"

Let's now hope that a committer gets around to consider these patch some 
day.


--
Fabien.


--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Raúl Marín Rodríguez 
Hi,


Indeed, this is quite strange...


 I don't want to go too deep into it, but you get stuff like this:

Select pow(2.0, -3)::text = pow(2, -3)::text;
 ?column?
--
 f
(1 row)


 - you can simplify the ipow function by removing handling of y<0 case,
>maybe add an assert to be sure to avoid it.

I agree, done.

 - you should add more symmetry and simplify the evaluation:

Done too.

 Add a test case to show what happens on NULL arguments, hopefully the
> result is NULL.

Done and it does.

Thanks again for the review.

On Mon, Nov 6, 2017 at 4:14 PM, Fabien COELHO  wrote:

>
> Hello,
>
> Sorry for the confusion, I wasn't aware that SQL pow changed types
>> depending on the input value.
>>
>
> Indeed, this is quite strange...
>
>   fabien=# SELECT i, POW(2, i) FROM generate_series(-2, 2) AS i;
>-2 | 0.25
>-1 | 0.5
> 0 | 1
> 1 | 2
> 2 | 4
>
> I've modified the function to match more closely the behaviour of SQL,
>> except that 0^(negative) returns 'double inf'. Do you think there is any
>> value in raising an error instead?
>>
>
>   fabien=# SELECT POW(0,-1);
>   ERROR:  zero raised to a negative power is undefined
>
> H... I'm fine with double inf, because exception in pgbench means the
> end of the script, which is not desirable for benchmarking purposes.
>
> I think that:
>
>  - you can simplify the ipow function by removing handling of y<0 case,
>maybe add an assert to be sure to avoid it.
>
>  - you should add more symmetry and simplify the evaluation:
>
>if (int & int)
>{
>   i1, i2 = ...;
>   if (i2 >= 0)
> setIntValue(retval, ipow(i1, i2));
>   else
> // conversion is done by C, no need to coerce again
> setDoubleValue(retval, pow(i1, i2));
>}
>else
>{
>  d1, d2 = ...;
>  setDoubleValue(retval, pow(d1, d2));
>}
>
> Add a test case to show what happens on NULL arguments, hopefully the
> result is NULL.
>
> --
> Fabien.
>



-- 

*Raúl Marín Rodríguez *carto.com
From 47f6ec396d3bc11c39066dbd4b31e75b76102094 Mon Sep 17 00:00:00 2001
From: Raul Marin 
Date: Fri, 13 Oct 2017 17:42:23 +0200
Subject: [PATCH] Add pow() support to pgbench

---
 doc/src/sgml/ref/pgbench.sgml|  7 
 src/bin/pgbench/exprparse.y  |  3 ++
 src/bin/pgbench/pgbench.c| 62 
 src/bin/pgbench/pgbench.h|  3 +-
 src/bin/pgbench/t/001_pgbench_with_server.pl | 15 +++
 5 files changed, 89 insertions(+), 1 deletion(-)

diff --git a/doc/src/sgml/ref/pgbench.sgml b/doc/src/sgml/ref/pgbench.sgml
index 1f55967e40a..32c94ba0dc1 100644
--- a/doc/src/sgml/ref/pgbench.sgml
+++ b/doc/src/sgml/ref/pgbench.sgml
@@ -1233,6 +1233,13 @@ pgbench  options  d
sqrt(2.0)
1.414213562
   
+  
+   pow(x, y)
+   integer if x and y are integers and y >= 0, else double
+   Numeric exponentiation
+   pow(2.0, 10)
+   1024.0
+  
  
  

diff --git a/src/bin/pgbench/exprparse.y b/src/bin/pgbench/exprparse.y
index 770be981f06..290bca99d12 100644
--- a/src/bin/pgbench/exprparse.y
+++ b/src/bin/pgbench/exprparse.y
@@ -334,6 +334,9 @@ static const struct
 	{
 		"!case_end", -2, PGBENCH_CASE
 	},
+	{
+		"pow", 2, PGBENCH_POW
+	},
 	/* keep as last array element */
 	{
 		NULL, 0, 0
diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index add653bf90c..3781e75721d 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -739,6 +739,27 @@ getPoissonRand(TState *thread, int64 center)
 }
 
 /*
+ * pow() for integer values with exp >= 0. Matches SQL pow() behaviour
+ */
+static int64
+ipow(int64 base, int64 exp)
+{
+	int64 result = 1;
+
+	Assert(exp >= 0);
+
+	while (exp)
+	{
+		if (exp & 1)
+			result *= base;
+		exp >>= 1;
+		base *= base;
+	}
+
+	return result;
+}
+
+/*
  * Initialize the given SimpleStats struct to all zeroes
  */
 static void
@@ -1918,6 +1939,47 @@ evalFunc(TState *thread, CState *st,
 return true;
 			}
 
+		case PGBENCH_POW:
+			{
+PgBenchValue *lval = &vargs[0];
+PgBenchValue *rval = &vargs[1];
+
+Assert(nargs == 2);
+
+/*
+ * If both operands are int and exp >= 0 use
+ * the ipow() function, else use pow()
+ */
+if (lval->type == PGBT_INT &&
+	 rval->type == PGBT_INT)
+{
+
+	int64		li,
+ri;
+
+	if (!coerceToInt(lval, &li) ||
+		!coerceToInt(rval, &ri))
+		return false;
+
+	if (ri >= 0)
+		setIntValue(retval, ipow(li, ri));
+	else
+		setDoubleValue(retval, pow(li, ri));
+}
+else
+{
+	double		ld,
+rd;
+
+	if (!coerceToDouble(lval, &ld) ||
+		!coerceToDouble(rval, &rd))
+		return false;
+
+	setDoubleValue(retval, pow(ld, rd));
+}
+return true;
+			}
+
 		default:
 			/* cannot get here */
 			Assert(0);
diff --git a/src/bin/pgbench/pgbench.h b/src/bin/pgbench/pgbench.h
index e1277a1dde6..9f26af92bf6 100644
--- a/src/bin/pgben

Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Fabien COELHO 


Hello,

Sorry for the confusion, I wasn't aware that SQL pow changed types 
depending on the input value.


Indeed, this is quite strange...

  fabien=# SELECT i, POW(2, i) FROM generate_series(-2, 2) AS i;
   -2 | 0.25
   -1 | 0.5
0 | 1
1 | 2
2 | 4

I've modified the function to match more closely the behaviour of SQL, 
except that 0^(negative) returns 'double inf'. Do you think there is any 
value in raising an error instead?


  fabien=# SELECT POW(0,-1);
  ERROR:  zero raised to a negative power is undefined

H... I'm fine with double inf, because exception in pgbench means the 
end of the script, which is not desirable for benchmarking purposes.


I think that:

 - you can simplify the ipow function by removing handling of y<0 case,
   maybe add an assert to be sure to avoid it.

 - you should add more symmetry and simplify the evaluation:

   if (int & int)
   {
  i1, i2 = ...;
  if (i2 >= 0)
setIntValue(retval, ipow(i1, i2));
  else
// conversion is done by C, no need to coerce again
setDoubleValue(retval, pow(i1, i2));
   }
   else
   {
 d1, d2 = ...;
 setDoubleValue(retval, pow(d1, d2));
   }

Add a test case to show what happens on NULL arguments, hopefully the 
result is NULL.


--
Fabien.


--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Raúl Marín Rodríguez 
Hi Fabien,

Sorry for the confusion, I wasn't aware that SQL pow changed types
depending on
the input value.

I've modified the function to match more closely the behaviour of SQL,
except
that 0^(negative) returns 'double inf'. Do you think there is any value in
raising an error instead?


On Mon, Nov 6, 2017 at 2:12 PM, Fabien COELHO  wrote:

>
> Hello Raúl,
>
> I've fixed the documentation and added an ipow function that handles both
>> positive and negative ints, having 0^0 == 1 and 0^(negative) ==
>> PG_INT64_MAX
>> since that's what my glibc math.h pow() is returning.
>>
>
> From the comment:
>
>  * For exp < 0 return 0 except when the base is 1 or -1
>
> I think that it should do what POW does in psql, i.e.:
>
>  fabien=# SELECT POW(2, -2); # 0.25
>
> that is if exp < 0 the double version should be used, it should
> not return 0.
>
> Basically the idea is that the pgbench client-side version should behave
> the same as the SQL version.
>
> --
> Fabien.




-- 

*Raúl Marín Rodríguez*carto.com
From be2fedcae277f7eede621dcda66b15b08372ce63 Mon Sep 17 00:00:00 2001
From: Raul Marin 
Date: Fri, 13 Oct 2017 17:42:23 +0200
Subject: [PATCH] Add pow() support to pgbench

---
 doc/src/sgml/ref/pgbench.sgml|  7 +++
 src/bin/pgbench/exprparse.y  |  3 +
 src/bin/pgbench/pgbench.c| 84 
 src/bin/pgbench/pgbench.h|  3 +-
 src/bin/pgbench/t/001_pgbench_with_server.pl | 13 +
 5 files changed, 109 insertions(+), 1 deletion(-)

diff --git a/doc/src/sgml/ref/pgbench.sgml b/doc/src/sgml/ref/pgbench.sgml
index 1f55967e40a..32c94ba0dc1 100644
--- a/doc/src/sgml/ref/pgbench.sgml
+++ b/doc/src/sgml/ref/pgbench.sgml
@@ -1233,6 +1233,13 @@ pgbench  options  d
sqrt(2.0)
1.414213562
   
+  
+   pow(x, y)
+   integer if x and y are integers and y >= 0, else double
+   Numeric exponentiation
+   pow(2.0, 10)
+   1024.0
+  
  
  

diff --git a/src/bin/pgbench/exprparse.y b/src/bin/pgbench/exprparse.y
index 770be981f06..290bca99d12 100644
--- a/src/bin/pgbench/exprparse.y
+++ b/src/bin/pgbench/exprparse.y
@@ -334,6 +334,9 @@ static const struct
 	{
 		"!case_end", -2, PGBENCH_CASE
 	},
+	{
+		"pow", 2, PGBENCH_POW
+	},
 	/* keep as last array element */
 	{
 		NULL, 0, 0
diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index add653bf90c..d565880b6e2 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -739,6 +739,51 @@ getPoissonRand(TState *thread, int64 center)
 }
 
 /*
+ * pow() for integer values
+ */
+static int64
+ipow(int64 base, int64 exp)
+{
+	int64 result;
+
+	if (base == 0)
+	{
+		if (exp > 0)
+			return 0;
+		else if (exp == 0)
+			return 1;
+		return PG_INT64_MAX;
+	}
+
+	/*
+	 * For exp > 0 calculate normally
+	 * For exp == 0 return 1 * sign of base
+	 * For exp < 0 return 0 except when the base is 1 or -1
+	 */
+	if (exp > 0)
+	{
+		result = 1;
+		while (exp)
+		{
+			if (exp & 1)
+result *= base;
+			exp >>= 1;
+			base *= base;
+		}
+	}
+	else if (exp == 0)
+		result = (base > 0) - (base < 0);
+	else
+	{
+		result = 1 / base;
+		if (exp % 2 == 0)
+			result *= result;
+	}
+
+	return result;
+}
+
+/*
  * Initialize the given SimpleStats struct to all zeroes
  */
 static void
@@ -1918,6 +1963,45 @@ evalFunc(TState *thread, CState *st,
 return true;
 			}
 
+		case PGBENCH_POW:
+			{
+PgBenchValue *lval = &vargs[0];
+PgBenchValue *rval = &vargs[1];
+double		ld,
+			rd;
+
+Assert(nargs == 2);
+
+/*
+ * If both operands are int and exp >= 0 use
+ * the ipow() function, else use pow()
+ */
+if (lval->type == PGBT_INT &&
+	 rval->type == PGBT_INT)
+{
+
+	int64		li,
+ri;
+
+	if (!coerceToInt(lval, &li) ||
+		!coerceToInt(rval, &ri))
+		return false;
+
+	if (ri >= 0)
+	{
+		setIntValue(retval, ipow(li, ri));
+		return true;
+	}
+}
+
+if (!coerceToDouble(lval, &ld) ||
+	!coerceToDouble(rval, &rd))
+	return false;
+
+setDoubleValue(retval, pow(ld, rd));
+return true;
+			}
+
 		default:
 			/* cannot get here */
 			Assert(0);
diff --git a/src/bin/pgbench/pgbench.h b/src/bin/pgbench/pgbench.h
index e1277a1dde6..9f26af92bf6 100644
--- a/src/bin/pgbench/pgbench.h
+++ b/src/bin/pgbench/pgbench.h
@@ -94,7 +94,8 @@ typedef enum PgBenchFunction
 	PGBENCH_LE,
 	PGBENCH_LT,
 	PGBENCH_IS,
-	PGBENCH_CASE
+	PGBENCH_CASE,
+	PGBENCH_POW
 } PgBenchFunction;
 
 typedef struct PgBenchExpr PgBenchExpr;
diff --git a/src/bin/pgbench/t/001_pgbench_with_server.pl b/src/bin/pgbench/t/001_pgbench_with_server.pl
index 8e19bbd3f45..2a1bb1216d7 100644
--- a/src/bin/pgbench/t/001_pgbench_with_server.pl
+++ b/src/bin/pgbench/t/001_pgbench_with_server.pl
@@ -237,6 +237,12 @@ sub pgbench
 		qr{command=36.: int 36\b},
 		qr{command=37.: boolean true\b},
 		qr{command=38.: boolean true\b},
+		qr{command=44.: int -27\b},
+		qr{command=45

Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Raúl Marín Rodríguez 
Hi Fabien,

Thanks for the review.
I've fixed the documentation and added an ipow function that handles both
positive and negative ints, having 0^0 == 1 and 0^(negative) == PG_INT64_MAX
since that's what my glibc math.h pow() is returning.

On Sat, Nov 4, 2017 at 12:34 PM, Fabien COELHO  wrote:

>
> Hello Raúl,
>
> Sorry about the patch. Attaching it now so it can be considered as
>> submitted.
>>
>
> There is a typo in the XML doc:
>
> 1024.0/
>
> Please check that the documentation compiles.
>
> I'm at odds with having the integer version rely on a double pow(), even
> if it works. I think that there should be a specific integer version which
> does use integer operations. From stack overflow, the following is
> suggested:
>
>  int ipow(int base, int exp)
>  {
> int result = 1;
> while (exp)
> {
> if (exp & 1)
> result *= base;
> exp >>= 1;
> base *= base;
> }
>
> return result;
>  }
>
> The integer version should be when x & y are integers *AND* y >= 0.
>
> if y is a negative integer, the double version should be used.
>
> --
> Fabien.




-- 

*Raúl Marín Rodríguez*carto.com
From 09105f8108e439834510ee5fb53036473f2977d5 Mon Sep 17 00:00:00 2001
From: Raul Marin 
Date: Fri, 13 Oct 2017 17:42:23 +0200
Subject: [PATCH] Add pow() support to pgbench

---
 doc/src/sgml/ref/pgbench.sgml|  7 
 src/bin/pgbench/exprparse.y  |  3 ++
 src/bin/pgbench/pgbench.c| 54 
 src/bin/pgbench/pgbench.h|  3 +-
 src/bin/pgbench/t/001_pgbench_with_server.pl |  9 +
 5 files changed, 75 insertions(+), 1 deletion(-)

diff --git a/doc/src/sgml/ref/pgbench.sgml b/doc/src/sgml/ref/pgbench.sgml
index 1f55967e40a..4e7f75ddf87 100644
--- a/doc/src/sgml/ref/pgbench.sgml
+++ b/doc/src/sgml/ref/pgbench.sgml
@@ -1233,6 +1233,13 @@ pgbench  options  d
sqrt(2.0)
1.414213562
   
+  
+   pow(x, y)
+   double if x or y are doubles, else integer
+   Numeric exponentiation
+   pow(2.0, 10)
+   1024.0
+  
  
  

diff --git a/src/bin/pgbench/exprparse.y b/src/bin/pgbench/exprparse.y
index 770be981f06..290bca99d12 100644
--- a/src/bin/pgbench/exprparse.y
+++ b/src/bin/pgbench/exprparse.y
@@ -334,6 +334,9 @@ static const struct
 	{
 		"!case_end", -2, PGBENCH_CASE
 	},
+	{
+		"pow", 2, PGBENCH_POW
+	},
 	/* keep as last array element */
 	{
 		NULL, 0, 0
diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index add653bf90c..bdf3e97682a 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -739,6 +739,51 @@ getPoissonRand(TState *thread, int64 center)
 }
 
 /*
+ * pow() for integer values
+ */
+static int64
+ipow(int64 base, int64 exp)
+{
+	int64 result;
+
+	if (base == 0)
+	{
+		if (exp > 0)
+			return 0;
+		else if (exp == 0)
+			return 1;
+		return PG_INT64_MAX;
+	}
+
+	/*
+	 * For exp > 0 calculate normally
+	 * For exp == 0 return 1 * sign of base
+	 * For exp < 0 return 0 except when the base is 1 or -1
+	 */
+	if (exp > 0)
+	{
+		result = 1;
+		while (exp)
+		{
+			if (exp & 1)
+result *= base;
+			exp >>= 1;
+			base *= base;
+		}
+	}
+	else if (exp == 0)
+		result = (base > 0) - (base < 0);
+	else
+	{
+		result = 1 / base;
+		if (exp % 2 == 0)
+			result *= result;
+	}
+
+	return result;
+}
+
+/*
  * Initialize the given SimpleStats struct to all zeroes
  */
 static void
@@ -1474,6 +1519,7 @@ evalFunc(TState *thread, CState *st,
 		case PGBENCH_NE:
 		case PGBENCH_LE:
 		case PGBENCH_LT:
+		case PGBENCH_POW:
 			{
 PgBenchValue *lval = &vargs[0],
 		   *rval = &vargs[1];
@@ -1525,6 +1571,10 @@ evalFunc(TState *thread, CState *st,
 			setBoolValue(retval, ld < rd);
 			return true;
 
+		case PGBENCH_POW:
+			setDoubleValue(retval, pow(ld, rd));
+			return true;
+
 		default:
 			/* cannot get here */
 			Assert(0);
@@ -1602,6 +1652,10 @@ evalFunc(TState *thread, CState *st,
 
 			return true;
 
+		case PGBENCH_POW:
+			setIntValue(retval, ipow(li, ri));
+			return true;
+
 		default:
 			/* cannot get here */
 			Assert(0);
diff --git a/src/bin/pgbench/pgbench.h b/src/bin/pgbench/pgbench.h
index e1277a1dde6..9f26af92bf6 100644
--- a/src/bin/pgbench/pgbench.h
+++ b/src/bin/pgbench/pgbench.h
@@ -94,7 +94,8 @@ typedef enum PgBenchFunction
 	PGBENCH_LE,
 	PGBENCH_LT,
 	PGBENCH_IS,
-	PGBENCH_CASE
+	PGBENCH_CASE,
+	PGBENCH_POW
 } PgBenchFunction;
 
 typedef struct PgBenchExpr PgBenchExpr;
diff --git a/src/bin/pgbench/t/001_pgbench_with_server.pl b/src/bin/pgbench/t/001_pgbench_with_server.pl
index 8e19bbd3f45..c5ecd749c40 100644
--- a/src/bin/pgbench/t/001_pgbench_with_server.pl
+++ b/src/bin/pgbench/t/001_pgbench_with_server.pl
@@ -237,6 +237,10 @@ sub pgbench
 		qr{command=36.: int 36\b},
 		qr{command=37.: boolean true\b},
 		qr{command=38.: boolean true\b},
+		qr{command=44.: int -27\b},
+		qr{command=45.: double

Re: [HACKERS] pow support for pgbench

2017-11-06 Thread Fabien COELHO 


Hello Raúl,


I've fixed the documentation and added an ipow function that handles both
positive and negative ints, having 0^0 == 1 and 0^(negative) == PG_INT64_MAX
since that's what my glibc math.h pow() is returning.



From the comment:


 * For exp < 0 return 0 except when the base is 1 or -1

I think that it should do what POW does in psql, i.e.:

 fabien=# SELECT POW(2, -2); # 0.25

that is if exp < 0 the double version should be used, it should
not return 0.

Basically the idea is that the pgbench client-side version should behave 
the same as the SQL version.


--
Fabien.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-11-04 Thread Fabien COELHO 


Hello Raúl,


Sorry about the patch. Attaching it now so it can be considered as
submitted.


There is a typo in the XML doc:

1024.0/

Please check that the documentation compiles.

I'm at odds with having the integer version rely on a double pow(), even 
if it works. I think that there should be a specific integer version which 
does use integer operations. From stack overflow, the following is 
suggested:


 int ipow(int base, int exp)
 {
int result = 1;
while (exp)
{
if (exp & 1)
result *= base;
exp >>= 1;
base *= base;
}

return result;
 }

The integer version should be when x & y are integers *AND* y >= 0.

if y is a negative integer, the double version should be used.

--
Fabien.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Fabien COELHO 


Hello Michaël,

I'm fine with having pow in pgbench.


I am adding as well Fabien in CC who worked in getting the internal
function infrastructure in the shape it is now (waaay better) with
commit 86c43f4.


It might be even better if https://commitfest.postgresql.org/15/985/, 
which has been around for over one year (!), get through some day...


--
Fabien.
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Raúl Marín Rodríguez 
Sorry about the patch. Attaching it now so it can be considered as
submitted.

-- 

*Raúl Marín Rodríguez*carto.com
From a6eecfe6637bdb0cbb02a9eda7b88d9c4325bb51 Mon Sep 17 00:00:00 2001
From: Raul Marin 
Date: Fri, 13 Oct 2017 17:42:23 +0200
Subject: [PATCH] Add pow() support to pgbench

---
 doc/src/sgml/ref/pgbench.sgml| 7 +++
 src/bin/pgbench/exprparse.y  | 3 +++
 src/bin/pgbench/pgbench.c| 9 +
 src/bin/pgbench/pgbench.h| 3 ++-
 src/bin/pgbench/t/001_pgbench_with_server.pl | 5 +
 5 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/doc/src/sgml/ref/pgbench.sgml b/doc/src/sgml/ref/pgbench.sgml
index 1f55967e40a..2e913dfcfd6 100644
--- a/doc/src/sgml/ref/pgbench.sgml
+++ b/doc/src/sgml/ref/pgbench.sgml
@@ -1233,6 +1233,13 @@ pgbench  options  d
sqrt(2.0)
1.414213562
   
+  
+   pow(x, y)
+   double if x or y are doubles, else integer
+   Numeric exponentiation
+   pow(2.0, 10)
+   1024.0/
+  
  
  

diff --git a/src/bin/pgbench/exprparse.y b/src/bin/pgbench/exprparse.y
index 770be981f06..290bca99d12 100644
--- a/src/bin/pgbench/exprparse.y
+++ b/src/bin/pgbench/exprparse.y
@@ -334,6 +334,9 @@ static const struct
 	{
 		"!case_end", -2, PGBENCH_CASE
 	},
+	{
+		"pow", 2, PGBENCH_POW
+	},
 	/* keep as last array element */
 	{
 		NULL, 0, 0
diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index add653bf90c..b2bab9b8239 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -1474,6 +1474,7 @@ evalFunc(TState *thread, CState *st,
 		case PGBENCH_NE:
 		case PGBENCH_LE:
 		case PGBENCH_LT:
+		case PGBENCH_POW:
 			{
 PgBenchValue *lval = &vargs[0],
 		   *rval = &vargs[1];
@@ -1525,6 +1526,10 @@ evalFunc(TState *thread, CState *st,
 			setBoolValue(retval, ld < rd);
 			return true;
 
+		case PGBENCH_POW:
+			setDoubleValue(retval, pow(ld, rd));
+			return true;
+
 		default:
 			/* cannot get here */
 			Assert(0);
@@ -1602,6 +1607,10 @@ evalFunc(TState *thread, CState *st,
 
 			return true;
 
+		case PGBENCH_POW:
+			setIntValue(retval, pow(li, ri));
+			return true;
+
 		default:
 			/* cannot get here */
 			Assert(0);
diff --git a/src/bin/pgbench/pgbench.h b/src/bin/pgbench/pgbench.h
index e1277a1dde6..9f26af92bf6 100644
--- a/src/bin/pgbench/pgbench.h
+++ b/src/bin/pgbench/pgbench.h
@@ -94,7 +94,8 @@ typedef enum PgBenchFunction
 	PGBENCH_LE,
 	PGBENCH_LT,
 	PGBENCH_IS,
-	PGBENCH_CASE
+	PGBENCH_CASE,
+	PGBENCH_POW
 } PgBenchFunction;
 
 typedef struct PgBenchExpr PgBenchExpr;
diff --git a/src/bin/pgbench/t/001_pgbench_with_server.pl b/src/bin/pgbench/t/001_pgbench_with_server.pl
index 8e19bbd3f45..81b7ce4dfcc 100644
--- a/src/bin/pgbench/t/001_pgbench_with_server.pl
+++ b/src/bin/pgbench/t/001_pgbench_with_server.pl
@@ -237,6 +237,8 @@ sub pgbench
 		qr{command=36.: int 36\b},
 		qr{command=37.: boolean true\b},
 		qr{command=38.: boolean true\b},
+		qr{command=44.: int -27\b},
+		qr{command=45.: double 1024\b},
 	],
 	'pgbench expressions',
 	{   '001_pgbench_expressions' => q{-- integer functions
@@ -299,6 +301,9 @@ sub pgbench
 \set v2 5432
 \set v3 -54.21E-2
 SELECT :v0, :v1, :v2, :v3;
+--- pow() operator
+\set poweri debug(pow(-3,3))
+\set powerd debug(pow(2.0,10))
 } });
 
 =head

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Raúl Marín Rodríguez 
Hi,

both patches seem complementary. I've rebased my changes on top of that
patch
(v14) in https://git.io/vFtnT and everything seems to be working fine.

On Mon, Oct 30, 2017 at 12:36 PM, Michael Paquier  wrote:

> On Mon, Oct 30, 2017 at 11:07 AM, Alvaro Herrera
>  wrote:
> > Michael Paquier wrote:
> >
> >> Please add this patch to the upcoming commit fest if you would like to
> >> get some feedback:
> >> https://commitfest.postgresql.org/15/
> >>
> >> I am adding as well Fabien in CC who worked in getting the internal
> >> function infrastructure in the shape it is now (waaay better) with
> >> commit 86c43f4.
> >
> > I think Raúl would do well to review this patch by Fabien
> > https://www.postgresql.org/message-id/alpine.DEB.2.20.
> 1710201835390.15170@lancre
> > which adds a few functions and operators.
>
> Good idea. pow() is not added by Fabien's patch, but an operator for
> pow() could be something to add as well.
> --
> Michael
>



-- 

*Raúl Marín Rodríguez*carto.com

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Alvaro Herrera 
Michael Paquier wrote:

> Attaching patches directly to a thread is a better practice as if
> github goes away, any Postgres developers can still have an access to
> any code you publish using the public archives on postgresql.org.

Also, by posting to pgsql-hackers indicating intention to integrate to
Postgres you're implicitly making a statement about the license of your
contribution; see the second half of the last paragraph at
https://wiki.postgresql.org/wiki/Archives_Policy

-- 
Álvaro Herrerahttps://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Michael Paquier 
On Mon, Oct 30, 2017 at 11:56 AM, Raúl Marín Rodríguez
 wrote:
> both patches seem complementary. I've rebased my changes on top of that
> patch
> (v14) in https://git.io/vFtnT and everything seems to be working fine.

Attaching patches directly to a thread is a better practice as if
github goes away, any Postgres developers can still have an access to
any code you publish using the public archives on postgresql.org.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Michael Paquier 
On Mon, Oct 30, 2017 at 11:07 AM, Alvaro Herrera
 wrote:
> Michael Paquier wrote:
>
>> Please add this patch to the upcoming commit fest if you would like to
>> get some feedback:
>> https://commitfest.postgresql.org/15/
>>
>> I am adding as well Fabien in CC who worked in getting the internal
>> function infrastructure in the shape it is now (waaay better) with
>> commit 86c43f4.
>
> I think Raúl would do well to review this patch by Fabien
> https://www.postgresql.org/message-id/alpine.DEB.2.20.1710201835390.15170@lancre
> which adds a few functions and operators.

Good idea. pow() is not added by Fabien's patch, but an operator for
pow() could be something to add as well.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Alvaro Herrera 
Michael Paquier wrote:

> Please add this patch to the upcoming commit fest if you would like to
> get some feedback:
> https://commitfest.postgresql.org/15/
> 
> I am adding as well Fabien in CC who worked in getting the internal
> function infrastructure in the shape it is now (waaay better) with
> commit 86c43f4.

I think Raúl would do well to review this patch by Fabien
https://www.postgresql.org/message-id/alpine.DEB.2.20.1710201835390.15170@lancre
which adds a few functions and operators.

-- 
Álvaro Herrerahttps://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Re: [HACKERS] pow support for pgbench

2017-10-30 Thread Michael Paquier 
On Fri, Oct 27, 2017 at 4:51 PM, Raúl Marín Rodríguez
 wrote:
> I've written a small patch to add support for pow() in pgbench.

Cool.

> The main reason behind it is that I'm currently using a shell call to do it
> which takes between 2-10 ms that can be a big percentage of the time taken
> by the whole transaction. For example (shortened):
>
> latency average = 11.718 ms
>  - statement latencies in milliseconds:
>  2.834  \setshell POWER2  awk 'BEGIN {p=2^ARGV[1]; print p }'
> :ZOOM_CURRENT
>  8.846  SELECT
> ST_AsBinary(ST_Simplify(ST_SnapToGrid("the_geom_webmercator",:SNAP),
> :SIMPLIFY)) AS geom FROM
>
> I've also updated the related docs and added some tests. Please let me know
> if I'm missing anything.

Please add this patch to the upcoming commit fest if you would like to
get some feedback:
https://commitfest.postgresql.org/15/

I am adding as well Fabien in CC who worked in getting the internal
function infrastructure in the shape it is now (waaay better) with
commit 86c43f4.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers