Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to content

Commit 20d6225

Browse files
tglsfdctglsfdc
committed
Add functions min_scale(numeric) and trim_scale(numeric).
These allow better control of trailing zeroes in numeric values. Pavel Stehule, based on an old proposal of Marko Tiikkaja's; review by Karl Pinc Discussion: https://postgr.es/m/CAFj8pRDjs-navGASeF0Wk74N36YGFJ+v=Ok9_knRa7vDc-qugg@mail.gmail.com
1 parent b9c130a commit 20d6225

File tree

6 files changed

+285
-3
lines changed

6 files changed

+285
-3
lines changed

‎doc/src/sgml/func.sgml

Lines changed: 30 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -918,6 +918,20 @@
918918
<entry><literal>6.0000000000</literal></entry>
919919
</row>
920920

921+
<row>
922+
<entry>
923+
<indexterm>
924+
<primary>min_scale</primary>
925+
</indexterm>
926+
<literal><function>min_scale(<type>numeric</type>)</function></literal>
927+
</entry>
928+
<entry><type>integer</type></entry>
929+
<entry>minimum scale (number of fractional decimal digits) needed
930+
to represent the supplied value</entry>
931+
<entry><literal>min_scale(8.4100)</literal></entry>
932+
<entry><literal>2</literal></entry>
933+
</row>
934+
921935
<row>
922936
<entry>
923937
<indexterm>
@@ -1011,8 +1025,8 @@
10111025
</entry>
10121026
<entry><type>integer</type></entry>
10131027
<entry>scale of the argument (the number of decimal digits in the fractional part)</entry>
1014-
<entry><literal>scale(8.41)</literal></entry>
1015-
<entry><literal>2</literal></entry>
1028+
<entry><literal>scale(8.4100)</literal></entry>
1029+
<entry><literal>4</literal></entry>
10161030
</row>
10171031

10181032
<row>
@@ -1041,6 +1055,20 @@
10411055
<entry><literal>1.4142135623731</literal></entry>
10421056
</row>
10431057

1058+
<row>
1059+
<entry>
1060+
<indexterm>
1061+
<primary>trim_scale</primary>
1062+
</indexterm>
1063+
<literal><function>trim_scale(<type>numeric</type>)</function></literal>
1064+
</entry>
1065+
<entry><type>numeric</type></entry>
1066+
<entry>reduce the scale (number of fractional decimal digits) by
1067+
removing trailing zeroes</entry>
1068+
<entry><literal>trim_scale(8.4100)</literal></entry>
1069+
<entry><literal>8.41</literal></entry>
1070+
</row>
1071+
10441072
<row>
10451073
<entry>
10461074
<indexterm>

‎src/backend/utils/adt/numeric.c

Lines changed: 91 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3179,6 +3179,97 @@ numeric_scale(PG_FUNCTION_ARGS)
31793179
PG_RETURN_INT32(NUMERIC_DSCALE(num));
31803180
}
31813181

3182+
/*
3183+
* Calculate minimum scale for value.
3184+
*/
3185+
static int
3186+
get_min_scale(NumericVar *var)
3187+
{
3188+
int min_scale;
3189+
int last_digit_pos;
3190+
3191+
/*
3192+
* Ordinarily, the input value will be "stripped" so that the last
3193+
* NumericDigit is nonzero. But we don't want to get into an infinite
3194+
* loop if it isn't, so explicitly find the last nonzero digit.
3195+
*/
3196+
last_digit_pos = var->ndigits - 1;
3197+
while (last_digit_pos >= 0 &&
3198+
var->digits[last_digit_pos] == 0)
3199+
last_digit_pos--;
3200+
3201+
if (last_digit_pos >= 0)
3202+
{
3203+
/* compute min_scale assuming that last ndigit has no zeroes */
3204+
min_scale = (last_digit_pos - var->weight) * DEC_DIGITS;
3205+
3206+
/*
3207+
* We could get a negative result if there are no digits after the
3208+
* decimal point. In this case the min_scale must be zero.
3209+
*/
3210+
if (min_scale > 0)
3211+
{
3212+
/*
3213+
* Reduce min_scale if trailing digit(s) in last NumericDigit are
3214+
* zero.
3215+
*/
3216+
NumericDigit last_digit = var->digits[last_digit_pos];
3217+
3218+
while (last_digit % 10 == 0)
3219+
{
3220+
min_scale--;
3221+
last_digit /= 10;
3222+
}
3223+
}
3224+
else
3225+
min_scale = 0;
3226+
}
3227+
else
3228+
min_scale = 0; /* result if input is zero */
3229+
3230+
return min_scale;
3231+
}
3232+
3233+
/*
3234+
* Returns minimum scale required to represent supplied value without loss.
3235+
*/
3236+
Datum
3237+
numeric_min_scale(PG_FUNCTION_ARGS)
3238+
{
3239+
Numeric num = PG_GETARG_NUMERIC(0);
3240+
NumericVar arg;
3241+
int min_scale;
3242+
3243+
if (NUMERIC_IS_NAN(num))
3244+
PG_RETURN_NULL();
3245+
3246+
init_var_from_num(num, &arg);
3247+
min_scale = get_min_scale(&arg);
3248+
free_var(&arg);
3249+
3250+
PG_RETURN_INT32(min_scale);
3251+
}
3252+
3253+
/*
3254+
* Reduce scale of numeric value to represent supplied value without loss.
3255+
*/
3256+
Datum
3257+
numeric_trim_scale(PG_FUNCTION_ARGS)
3258+
{
3259+
Numeric num = PG_GETARG_NUMERIC(0);
3260+
Numeric res;
3261+
NumericVar result;
3262+
3263+
if (NUMERIC_IS_NAN(num))
3264+
PG_RETURN_NUMERIC(make_result(&const_nan));
3265+
3266+
init_var_from_num(num, &result);
3267+
result.dscale = get_min_scale(&result);
3268+
res = make_result(&result);
3269+
free_var(&result);
3270+
3271+
PG_RETURN_NUMERIC(res);
3272+
}
31823273

31833274

31843275
/* ----------------------------------------------------------------------

‎src/include/catalog/catversion.h

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -53,6 +53,6 @@
5353
*/
5454

5555
/* yyyymmddN */
56-
#define CATALOG_VERSION_NO 201911241
56+
#define CATALOG_VERSION_NO 202001061
5757

5858
#endif

‎src/include/catalog/pg_proc.dat

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4254,6 +4254,13 @@
42544254
{ oid => '3281', descr => 'number of decimal digits in the fractional part',
42554255
proname => 'scale', prorettype => 'int4', proargtypes => 'numeric',
42564256
prosrc => 'numeric_scale' },
4257+
{ oid => '8389', descr => 'minimum scale needed to represent the value',
4258+
proname => 'min_scale', prorettype => 'int4', proargtypes => 'numeric',
4259+
prosrc => 'numeric_min_scale' },
4260+
{ oid => '8390',
4261+
descr => 'numeric with minimum scale needed to represent the value',
4262+
proname => 'trim_scale', prorettype => 'numeric', proargtypes => 'numeric',
4263+
prosrc => 'numeric_trim_scale' },
42574264
{ oid => '1740', descr => 'convert int4 to numeric',
42584265
proname => 'numeric', prorettype => 'numeric', proargtypes => 'int4',
42594266
prosrc => 'int4_numeric' },

‎src/test/regress/expected/numeric.out

Lines changed: 126 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2078,6 +2078,132 @@ select scale(-13.000000000000000);
20782078
15
20792079
(1 row)
20802080

2081+
--
2082+
-- Tests for min_scale()
2083+
--
2084+
select min_scale(numeric 'NaN') is NULL; -- should be true
2085+
?column?
2086+
----------
2087+
t
2088+
(1 row)
2089+
2090+
select min_scale(0); -- no digits
2091+
min_scale
2092+
-----------
2093+
0
2094+
(1 row)
2095+
2096+
select min_scale(0.00); -- no digits again
2097+
min_scale
2098+
-----------
2099+
0
2100+
(1 row)
2101+
2102+
select min_scale(1.0); -- no scale
2103+
min_scale
2104+
-----------
2105+
0
2106+
(1 row)
2107+
2108+
select min_scale(1.1); -- scale 1
2109+
min_scale
2110+
-----------
2111+
1
2112+
(1 row)
2113+
2114+
select min_scale(1.12); -- scale 2
2115+
min_scale
2116+
-----------
2117+
2
2118+
(1 row)
2119+
2120+
select min_scale(1.123); -- scale 3
2121+
min_scale
2122+
-----------
2123+
3
2124+
(1 row)
2125+
2126+
select min_scale(1.1234); -- scale 4, filled digit
2127+
min_scale
2128+
-----------
2129+
4
2130+
(1 row)
2131+
2132+
select min_scale(1.12345); -- scale 5, 2 NDIGITS
2133+
min_scale
2134+
-----------
2135+
5
2136+
(1 row)
2137+
2138+
select min_scale(1.1000); -- 1 pos in NDIGITS
2139+
min_scale
2140+
-----------
2141+
1
2142+
(1 row)
2143+
2144+
select min_scale(1e100); -- very big number
2145+
min_scale
2146+
-----------
2147+
0
2148+
(1 row)
2149+
2150+
--
2151+
-- Tests for trim_scale()
2152+
--
2153+
select trim_scale(numeric 'NaN');
2154+
trim_scale
2155+
------------
2156+
NaN
2157+
(1 row)
2158+
2159+
select trim_scale(1.120);
2160+
trim_scale
2161+
------------
2162+
1.12
2163+
(1 row)
2164+
2165+
select trim_scale(0);
2166+
trim_scale
2167+
------------
2168+
0
2169+
(1 row)
2170+
2171+
select trim_scale(0.00);
2172+
trim_scale
2173+
------------
2174+
0
2175+
(1 row)
2176+
2177+
select trim_scale(1.1234500);
2178+
trim_scale
2179+
------------
2180+
1.12345
2181+
(1 row)
2182+
2183+
select trim_scale(110123.12475871856128000);
2184+
trim_scale
2185+
-----------------------
2186+
110123.12475871856128
2187+
(1 row)
2188+
2189+
select trim_scale(-1123.124718561280000000);
2190+
trim_scale
2191+
-------------------
2192+
-1123.12471856128
2193+
(1 row)
2194+
2195+
select trim_scale(-13.00000000000000000000);
2196+
trim_scale
2197+
------------
2198+
-13
2199+
(1 row)
2200+
2201+
select trim_scale(1e100);
2202+
trim_scale
2203+
-------------------------------------------------------------------------------------------------------
2204+
10000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
2205+
(1 row)
2206+
20812207
--
20822208
-- Tests for SUM()
20832209
--

‎src/test/regress/sql/numeric.sql

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1036,6 +1036,36 @@ select scale(110123.12475871856128);
10361036
select scale(-1123.12471856128);
10371037
select scale(-13.000000000000000);
10381038

1039+
--
1040+
-- Tests for min_scale()
1041+
--
1042+
1043+
select min_scale(numeric 'NaN') is NULL; -- should be true
1044+
select min_scale(0); -- no digits
1045+
select min_scale(0.00); -- no digits again
1046+
select min_scale(1.0); -- no scale
1047+
select min_scale(1.1); -- scale 1
1048+
select min_scale(1.12); -- scale 2
1049+
select min_scale(1.123); -- scale 3
1050+
select min_scale(1.1234); -- scale 4, filled digit
1051+
select min_scale(1.12345); -- scale 5, 2 NDIGITS
1052+
select min_scale(1.1000); -- 1 pos in NDIGITS
1053+
select min_scale(1e100); -- very big number
1054+
1055+
--
1056+
-- Tests for trim_scale()
1057+
--
1058+
1059+
select trim_scale(numeric 'NaN');
1060+
select trim_scale(1.120);
1061+
select trim_scale(0);
1062+
select trim_scale(0.00);
1063+
select trim_scale(1.1234500);
1064+
select trim_scale(110123.12475871856128000);
1065+
select trim_scale(-1123.124718561280000000);
1066+
select trim_scale(-13.00000000000000000000);
1067+
select trim_scale(1e100);
1068+
10391069
--
10401070
-- Tests for SUM()
10411071
--

0 commit comments

Comments
 (0)