2015-12-01 11:12 GMT+01:00 Pavel Stehule <pavel.steh...@gmail.com>:

>
>
> 2015-12-01 11:02 GMT+01:00 Kyotaro HORIGUCHI <
> horiguchi.kyot...@lab.ntt.co.jp>:
>
>> Hello, The meaning of "be orange" (I couldn't find it) interests
>> me but putting it aside..
>>
>> I have some random comments.
>>
>> At Mon, 30 Nov 2015 18:30:18 +0100, Pavel Stehule <
>> pavel.steh...@gmail.com> wrote in <
>> cafj8prcd6we3tqmr0vbcn98wf0p3o3h6nau4btaoswcj443...@mail.gmail.com>
>> > Hi
>> >
>> >
>> > > 2. using independent implementation - there is some redundant code,
>> but we
>> > > can support duble insted int, and we can support some additional
>> units.
>> > >
>> >
>> >  new patch is based on own implementation - use numeric/bigint
>> calculations
>> >
>> > + regress tests and doc
>>
>> 1. What do you think about binary byte prefixes? (kiB, MiB...)
>>
>
> I don't know this mechanics - can you write it? It can be good idea/
>
>
>>
>>  I couldn't read what Robert wrote upthread but I also would like
>>  to have 2-base byte unit prifixes. (Sorry I couldn't put it
>>  aside..)
>>
>>
>> 2. Doesn't it allow units in lower case?
>>
>
> The parser is consistent with a behave used in configure file processing.
> We can change it, but then there will be divergence between this function
> and GUC parser.
>
>
>>
>>  I think '1gb' and so should be acceptable as an input.
>>
>>
>> 3. Why are you allow positive sign prefixing digits? I suppose
>>   that positive sign also shouldn't be allowed if it rejects
>>   prifixing negative sign.
>>
>
> I have not strong opinion about it. '-' is exactly wrong, but '+' can be
> acceptable. But it can be changed.
>
>
>>
>> 4. Useless includes
>>
>>  dbsize.c is modified to include guc.h but it is useless.
>>
>
> I'll fix it.
>
>
>>
>> 5. Error message
>>
>> +       if (ndigits == 0)
>> +               ereport(ERROR,
>> +                               (errcode(ERRCODE_INVALID_PARAMETER_VALUE),
>> +                                        errmsg("string is empty")));
>>
>>  If pg_size_bytes allows prefixing positive sign or spaces,
>>  ndigits == 0 here doesn't mean that the whole string is empty.
>>
>>
> I'll fix it.
>
>
>>
>> 6. Symmetry with pg_size_pretty
>>
>>  pg_size_pretty doesn't have units larger than TB. Just adding PB
>>  to it breaks backward compatibility, but leaving as it is breaks
>>  symmetry with this function. Some possible solution for this
>>  that I can guess for now are..
>>
>
> I prefer a warning about possible compatibility issue in pg_size_pretty
> and support PB directly there.
>
>
>>
>>   - Leave it as is.
>>
>>   - New GUC to allow PB for pg_size_pretty().
>>
>>   - Expanded function such like pg_size_pretty2 (oops..)
>>
>>   - New polymorphic (sibling?) function with additional
>>     parameters to instruct how they work. (The following
>>     involving 2-base representations)
>>
>>     pg_size_pretty(n, <2base>, <allow_PB or such..>);
>>     pg_size_bytes(n, <2base>, <allow_PB or such..>);
>>
>> 7. Docs
>>
>> +          Converts a size in human-readable format with size units
>> +          to bytes
>>
>>  The descriptions for the functions around use 'into' instead of
>>  'to' for the preposition.
>>
>>
>> 8. Regression
>>
>>  The regression in the patch looks good enough as is (except that
>>  it forgets the unit 'B' or prifixing spaces or sings), but they
>>  are necessarily have individual tests. The following SQL
>>  statement will do them at once.
>>
>>   SELECT s, pg_size_bytes(s) FROM (VALUES ('1'), ('1B')...) as t(s);
>>
>>
> I'll fix it.
>

here is updated patch

Regards

Pavel


>
> Thank you for your ideas
>
> Regards
>
> Pavel
>
>
>> regards,
>>
>> --
>> Kyotaro Horiguchi
>> NTT Open Source Software Center
>>
>>
>>
>
diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
new file mode 100644
index 60b9a09..55e7d35
*** a/doc/src/sgml/func.sgml
--- b/doc/src/sgml/func.sgml
*************** postgres=# SELECT * FROM pg_xlogfile_nam
*** 17607,17612 ****
--- 17607,17615 ----
      <primary>pg_relation_size</primary>
     </indexterm>
     <indexterm>
+     <primary>pg_size_bytes</primary>
+    </indexterm>
+    <indexterm>
      <primary>pg_size_pretty</primary>
     </indexterm>
     <indexterm>
*************** postgres=# SELECT * FROM pg_xlogfile_nam
*** 17677,17682 ****
--- 17680,17695 ----
         </entry>
        </row>
        <row>
+        <entry>
+         <literal><function>pg_size_bytes(<type>text</type>)</function></literal>
+         </entry>
+        <entry><type>bigint</type></entry>
+        <entry>
+          Converts a size in human-readable format with size units
+          into bytes.
+        </entry>
+       </row>
+       <row>
         <entry>
          <literal><function>pg_size_pretty(<type>bigint</type>)</function></literal>
          </entry>
diff --git a/src/backend/utils/adt/dbsize.c b/src/backend/utils/adt/dbsize.c
new file mode 100644
index 5ee59d0..0e7399c
*** a/src/backend/utils/adt/dbsize.c
--- b/src/backend/utils/adt/dbsize.c
***************
*** 31,36 ****
--- 31,58 ----
  #include "utils/relmapper.h"
  #include "utils/syscache.h"
  
+ #define MAX_UNIT_LEN		3
+ #define MAX_DIGITS		20
+ 
+ typedef struct
+ {
+ 	char		unit[MAX_UNIT_LEN + 1];
+ 	long int	multiplier;
+ } unit_multiplier;
+ 
+ static const unit_multiplier unit_multiplier_table[] =
+ {
+ 	{"PB", 1024L * 1024 * 1024 * 1024 * 1024},
+ 	{"TB", 1024L * 1024 * 1024 * 1024},
+ 	{"GB", 1024L * 1024 * 1024},
+ 	{"MB", 1024L * 1024},
+ 	{"kB", 1024L},
+ 	{"B", 1L},
+ 
+ 	{""}				/* end of table marker */
+ };
+ 
+ 
  /* Divide by two and round towards positive infinity. */
  #define half_rounded(x)   (((x) + ((x) < 0 ? 0 : 1)) / 2)
  
*************** pg_size_pretty(PG_FUNCTION_ARGS)
*** 559,566 ****
  				else
  				{
  					size >>= 10;
! 					snprintf(buf, sizeof(buf), INT64_FORMAT " TB",
! 							 half_rounded(size));
  				}
  			}
  		}
--- 581,595 ----
  				else
  				{
  					size >>= 10;
! 					if (Abs(size) < limit2)
! 						snprintf(buf, sizeof(buf), INT64_FORMAT " TB",
! 								 half_rounded(size));
! 					else
! 					{
! 						size >>= 10;
! 						snprintf(buf, sizeof(buf), INT64_FORMAT " PB",
! 								 half_rounded(size));
! 					}
  				}
  			}
  		}
*************** pg_size_pretty_numeric(PG_FUNCTION_ARGS)
*** 689,696 ****
  				{
  					/* size >>= 10 */
  					size = numeric_shift_right(size, 10);
! 					size = numeric_half_rounded(size);
! 					result = psprintf("%s TB", numeric_to_cstring(size));
  				}
  			}
  		}
--- 718,736 ----
  				{
  					/* size >>= 10 */
  					size = numeric_shift_right(size, 10);
! 
! 					if (numeric_is_less(numeric_absolute(size), limit2))
! 					{
! 						size = numeric_half_rounded(size);
! 						result = psprintf("%s TB", numeric_to_cstring(size));
! 					}
! 					else
! 					{
! 						/* size >>= 10 */
! 						size = numeric_shift_right(size, 10);
! 						size = numeric_half_rounded(size);
! 						result = psprintf("%s PB", numeric_to_cstring(size));
! 					}
  				}
  			}
  		}
*************** pg_size_pretty_numeric(PG_FUNCTION_ARGS)
*** 700,705 ****
--- 740,848 ----
  }
  
  /*
+  * Convert human readable size to long int.
+  *
+  * cannot to use parse_int due too low limits there
+  */
+ Datum
+ pg_size_bytes(PG_FUNCTION_ARGS)
+ {
+ 	text	   *arg = PG_GETARG_TEXT_PP(0);
+ 	const char	   *cp = text_to_cstring(arg);
+ 	char		digits[MAX_DIGITS + 1];
+ 	int		ndigits = 0;
+ 	Numeric			num;
+ 	int64 result;
+ 
+ 	/* Skip leading spaces */
+ 	while (isspace((unsigned char) *cp))
+ 		cp++;
+ 
+ 	switch (*cp)
+ 	{
+ 		/* ignore plus symbol */
+ 		case '+':
+ 			cp++;
+ 			break;
+ 		case '-':
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("size cannot be negative")));
+ 	}
+ 
+ 	while (*cp)
+ 	{
+ 		if ((isdigit(*cp) || *cp == '.') && ndigits < MAX_DIGITS)
+ 			digits[ndigits++] = *cp++;
+ 		else
+ 			break;
+ 	}
+ 
+ 	if (ndigits == 0)
+ 		ereport(ERROR,
+ 				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("invalid format")));
+ 
+ 	digits[ndigits] = '\0';
+ 	num = DatumGetNumeric(DirectFunctionCall3(numeric_in,
+ 							CStringGetDatum(digits), 0, -1));
+ 
+ 	/* allow whitespace between integer and unit */
+ 	while (isspace((unsigned char) *cp))
+ 		cp++;
+ 
+ 	/* Handle possible unit */
+ 	if (*cp != '\0')
+ 	{
+ 		char		unit[MAX_UNIT_LEN + 1];
+ 		int		unitlen = 0;
+ 		int		i;
+ 		long int	multiplier;
+ 		bool			found = false;
+ 		Numeric			mul_num;
+ 
+ 
+ 		while (*cp && !isspace(*cp) && unitlen < MAX_UNIT_LEN)
+ 			unit[unitlen++] = *cp++;
+ 
+ 		unit[unitlen] = '\0';
+ 
+ 		/* allow whitespace after unit */
+ 		while (isspace((unsigned char) *cp))
+ 			cp++;
+ 
+ 		if (*cp != '\0')
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("invalid format")));
+ 
+ 		for (i = 0; *unit_multiplier_table[i].unit; i++)
+ 		{
+ 			if (strcmp(unit, unit_multiplier_table[i].unit) == 0)
+ 			{
+ 				multiplier = unit_multiplier_table[i].multiplier;
+ 				found = true;
+ 				break;
+ 			}
+ 		}
+ 
+ 		if (!found)
+ 			ereport(ERROR,
+ 					(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
+ 					 errmsg("unknown unit \"%s\"", unit)));
+ 
+ 		mul_num = DatumGetNumeric(DirectFunctionCall1(int8_numeric, Int64GetDatum(multiplier)));
+ 		num = DatumGetNumeric(DirectFunctionCall2(numeric_mul,
+ 							    NumericGetDatum(mul_num),
+ 							    NumericGetDatum(num)));
+ 	}
+ 
+ 	result = DatumGetInt64(DirectFunctionCall1(numeric_int8, NumericGetDatum(num)));
+ 
+ 	PG_RETURN_INT64(result);
+ }
+ 
+ /*
   * Get the filenode of a relation
   *
   * This is expected to be used in queries like
diff --git a/src/include/catalog/pg_proc.h b/src/include/catalog/pg_proc.h
new file mode 100644
index d8640db..b68c8fa
*** a/src/include/catalog/pg_proc.h
--- b/src/include/catalog/pg_proc.h
*************** DATA(insert OID = 2286 ( pg_total_relati
*** 3662,3667 ****
--- 3662,3669 ----
  DESCR("total disk space usage for the specified table and associated indexes");
  DATA(insert OID = 2288 ( pg_size_pretty			PGNSP PGUID 12 1 0 0 0 f f f f t f v s 1 0 25 "20" _null_ _null_ _null_ _null_ _null_ pg_size_pretty _null_ _null_ _null_ ));
  DESCR("convert a long int to a human readable text using size units");
+ DATA(insert OID = 3317 ( pg_size_bytes			PGNSP PGUID 12 1 0 0 0 f f f f t f v s 1 0 20 "25" _null_ _null_ _null_ _null_ _null_ pg_size_bytes _null_ _null_ _null_ ));
+ DESCR("convert a human readable text with size units to long int bytes");
  DATA(insert OID = 3166 ( pg_size_pretty			PGNSP PGUID 12 1 0 0 0 f f f f t f v s 1 0 25 "1700" _null_ _null_ _null_ _null_ _null_ pg_size_pretty_numeric _null_ _null_ _null_ ));
  DESCR("convert a numeric to a human readable text using size units");
  DATA(insert OID = 2997 ( pg_table_size			PGNSP PGUID 12 1 0 0 0 f f f f t f v s 1 0 20 "2205" _null_ _null_ _null_ _null_ _null_ pg_table_size _null_ _null_ _null_ ));
diff --git a/src/include/utils/builtins.h b/src/include/utils/builtins.h
new file mode 100644
index e610bf3..227e5f5
*** a/src/include/utils/builtins.h
--- b/src/include/utils/builtins.h
*************** extern Datum pg_relation_size(PG_FUNCTIO
*** 462,467 ****
--- 462,468 ----
  extern Datum pg_total_relation_size(PG_FUNCTION_ARGS);
  extern Datum pg_size_pretty(PG_FUNCTION_ARGS);
  extern Datum pg_size_pretty_numeric(PG_FUNCTION_ARGS);
+ extern Datum pg_size_bytes(PG_FUNCTION_ARGS);
  extern Datum pg_table_size(PG_FUNCTION_ARGS);
  extern Datum pg_indexes_size(PG_FUNCTION_ARGS);
  extern Datum pg_relation_filenode(PG_FUNCTION_ARGS);
diff --git a/src/test/regress/expected/dbsize.out b/src/test/regress/expected/dbsize.out
new file mode 100644
index aa513e7..4ed9190
*** a/src/test/regress/expected/dbsize.out
--- b/src/test/regress/expected/dbsize.out
***************
*** 1,37 ****
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::bigint), (1000::bigint), (1000000::bigint),
              (1000000000::bigint), (1000000000000::bigint),
!             (1000000000000000::bigint)) x(size);
!        size       | pg_size_pretty | pg_size_pretty 
! ------------------+----------------+----------------
!                10 | 10 bytes       | -10 bytes
!              1000 | 1000 bytes     | -1000 bytes
!           1000000 | 977 kB         | -977 kB
!        1000000000 | 954 MB         | -954 MB
!     1000000000000 | 931 GB         | -931 GB
!  1000000000000000 | 909 TB         | -909 TB
! (6 rows)
  
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::numeric), (1000::numeric), (1000000::numeric),
              (1000000000::numeric), (1000000000000::numeric),
              (1000000000000000::numeric),
              (10.5::numeric), (1000.5::numeric), (1000000.5::numeric),
              (1000000000.5::numeric), (1000000000000.5::numeric),
!             (1000000000000000.5::numeric)) x(size);
!         size        | pg_size_pretty | pg_size_pretty 
! --------------------+----------------+----------------
!                  10 | 10 bytes       | -10 bytes
!                1000 | 1000 bytes     | -1000 bytes
!             1000000 | 977 kB         | -977 kB
!          1000000000 | 954 MB         | -954 MB
!       1000000000000 | 931 GB         | -931 GB
!    1000000000000000 | 909 TB         | -909 TB
!                10.5 | 10.5 bytes     | -10.5 bytes
!              1000.5 | 1000.5 bytes   | -1000.5 bytes
!           1000000.5 | 977 kB         | -977 kB
!        1000000000.5 | 954 MB         | -954 MB
!     1000000000000.5 | 931 GB         | -931 GB
!  1000000000000000.5 | 909 TB         | -909 TB
! (12 rows)
  
--- 1,63 ----
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::bigint), (1000::bigint), (1000000::bigint),
              (1000000000::bigint), (1000000000000::bigint),
!             (1000000000000000::bigint),
!             (1000000000000000000::bigint)) x(size);
!         size         | pg_size_pretty | pg_size_pretty 
! ---------------------+----------------+----------------
!                   10 | 10 bytes       | -10 bytes
!                 1000 | 1000 bytes     | -1000 bytes
!              1000000 | 977 kB         | -977 kB
!           1000000000 | 954 MB         | -954 MB
!        1000000000000 | 931 GB         | -931 GB
!     1000000000000000 | 909 TB         | -909 TB
!  1000000000000000000 | 888 PB         | -888 PB
! (7 rows)
  
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::numeric), (1000::numeric), (1000000::numeric),
              (1000000000::numeric), (1000000000000::numeric),
              (1000000000000000::numeric),
+             (1000000000000000000::numeric),
              (10.5::numeric), (1000.5::numeric), (1000000.5::numeric),
              (1000000000.5::numeric), (1000000000000.5::numeric),
!             (1000000000000000.5::numeric),
!             (1000000000000000000.5::numeric)) x(size);
!          size          | pg_size_pretty | pg_size_pretty 
! -----------------------+----------------+----------------
!                     10 | 10 bytes       | -10 bytes
!                   1000 | 1000 bytes     | -1000 bytes
!                1000000 | 977 kB         | -977 kB
!             1000000000 | 954 MB         | -954 MB
!          1000000000000 | 931 GB         | -931 GB
!       1000000000000000 | 909 TB         | -909 TB
!    1000000000000000000 | 888 PB         | -888 PB
!                   10.5 | 10.5 bytes     | -10.5 bytes
!                 1000.5 | 1000.5 bytes   | -1000.5 bytes
!              1000000.5 | 977 kB         | -977 kB
!           1000000000.5 | 954 MB         | -954 MB
!        1000000000000.5 | 931 GB         | -931 GB
!     1000000000000000.5 | 909 TB         | -909 TB
!  1000000000000000000.5 | 888 PB         | -888 PB
! (14 rows)
  
+ SELECT pg_size_bytes(size) FROM
+    (VALUES('1'), ('1B'), ('1kB'), ('1MB'), (' 1 GB'), ('1.5 GB '),
+           ('1TB'),('1PB')) x(size);
+   pg_size_bytes   
+ ------------------
+                 1
+                 1
+              1024
+           1048576
+        1073741824
+        1610612736
+     1099511627776
+  1125899906842624
+ (8 rows)
+ 
+ --should fail
+ SELECT pg_size_bytes('1 AB');
+ ERROR:  unknown unit "AB"
+ SELECT pg_size_bytes('1 AB A');
+ ERROR:  invalid format
diff --git a/src/test/regress/sql/dbsize.sql b/src/test/regress/sql/dbsize.sql
new file mode 100644
index c118090..6cf833c
*** a/src/test/regress/sql/dbsize.sql
--- b/src/test/regress/sql/dbsize.sql
***************
*** 1,12 ****
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::bigint), (1000::bigint), (1000000::bigint),
              (1000000000::bigint), (1000000000000::bigint),
!             (1000000000000000::bigint)) x(size);
  
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::numeric), (1000::numeric), (1000000::numeric),
              (1000000000::numeric), (1000000000000::numeric),
              (1000000000000000::numeric),
              (10.5::numeric), (1000.5::numeric), (1000000.5::numeric),
              (1000000000.5::numeric), (1000000000000.5::numeric),
!             (1000000000000000.5::numeric)) x(size);
--- 1,23 ----
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::bigint), (1000::bigint), (1000000::bigint),
              (1000000000::bigint), (1000000000000::bigint),
!             (1000000000000000::bigint),
!             (1000000000000000000::bigint)) x(size);
  
  SELECT size, pg_size_pretty(size), pg_size_pretty(-1 * size) FROM
      (VALUES (10::numeric), (1000::numeric), (1000000::numeric),
              (1000000000::numeric), (1000000000000::numeric),
              (1000000000000000::numeric),
+             (1000000000000000000::numeric),
              (10.5::numeric), (1000.5::numeric), (1000000.5::numeric),
              (1000000000.5::numeric), (1000000000000.5::numeric),
!             (1000000000000000.5::numeric),
!             (1000000000000000000.5::numeric)) x(size);
! 
! SELECT pg_size_bytes(size) FROM
!    (VALUES('1'), ('1B'), ('1kB'), ('1MB'), (' 1 GB'), ('1.5 GB '),
!           ('1TB'),('1PB')) x(size);
! 
! --should fail
! SELECT pg_size_bytes('1 AB');
! SELECT pg_size_bytes('1 AB A');
\ No newline at end of file
-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Reply via email to