Skip to content

Commit 0e13d62

Browse files
jconwayjconway
committed
Require that array literals produce "rectangular" arrays, i.e. all the
subarrays of a given dimension have the same number of elements/subarrays. Also repair a longstanding undocumented (as far as I can see) ability to explicitly set array bounds in the array literal syntax. It now can deal properly with negative array indicies. Modify array_out so that arrays with non-standard lower bounds (i.e. not 1) are output with the expicit dimension syntax. This fixes a longstanding issue whereby arrays with non-default lower bounds had them changed to default after a dump/reload cycle. Modify regression tests and docs to suit, and add some minimal documentation regarding the explicit dimension syntax.
1 parent 39ec59f commit 0e13d62

File tree

5 files changed

+193
-89
lines changed

5 files changed

+193
-89
lines changed

‎doc/src/sgml/array.sgml‎

Lines changed: 70 additions & 45 deletions
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
<!-- $PostgreSQL: pgsql/doc/src/sgml/array.sgml,v 1.35 2004/06/07 04:04:47 tgl Exp $ -->
1+
<!-- $PostgreSQL: pgsql/doc/src/sgml/array.sgml,v 1.36 2004/08/05 03:29:11 joe Exp $ -->
22

33
<sect1 id="arrays">
44
<title>Arrays</title>
@@ -121,12 +121,23 @@ CREATE TABLE tictactoe (
121121
INSERT INTO sal_emp
122122
VALUES ('Bill',
123123
'{10000, 10000, 10000, 10000}',
124-
'{{"meeting", "lunch"}, {}}');
124+
'{{"meeting", "lunch"}, {"meeting"}}');
125+
ERROR: multidimensional arrays must have array expressions with matching dimensions
126+
</programlisting>
127+
128+
Note that multidimensional arrays must have matching extents for each
129+
dimension. A mismatch causes an error report.
130+
131+
<programlisting>
132+
INSERT INTO sal_emp
133+
VALUES ('Bill',
134+
'{10000, 10000, 10000, 10000}',
135+
'{{"meeting", "lunch"}, {"training", "presentation"}}');
125136

126137
INSERT INTO sal_emp
127138
VALUES ('Carol',
128139
'{20000, 25000, 25000, 25000}',
129-
'{{"talk", "consult"}, {"meeting"}}');
140+
'{{"breakfast", "consulting"}, {"meeting", "lunch"}}');
130141
</programlisting>
131142
</para>
132143

@@ -138,59 +149,31 @@ INSERT INTO sal_emp
138149
</para>
139150

140151
<para>
141-
This can lead to surprising results. For example, the result of the
142-
previous two inserts looks like this:
152+
The result of the previous two inserts looks like this:
143153
<programlisting>
144154
SELECT * FROM sal_emp;
145-
name | pay_by_quarter | schedule
146-
-------+---------------------------+--------------------
147-
Bill | {10000,10000,10000,10000} | {{meeting},{""}}
148-
Carol | {20000,25000,25000,25000} | {{talk},{meeting}}
155+
name | pay_by_quarter | schedule
156+
-------+---------------------------+-------------------------------------------
157+
Bill | {10000,10000,10000,10000} | {{meeting,lunch},{training,presentation}}
158+
Carol | {20000,25000,25000,25000} | {{breakfast,consulting},{meeting,lunch}}
149159
(2 rows)
150160
</programlisting>
151-
Because the <literal>[2][2]</literal> element of
152-
<structfield>schedule</structfield> is missing in each of the
153-
<command>INSERT</command> statements, the <literal>[1][2]</literal>
154-
element is discarded.
155161
</para>
156162

157-
<note>
158-
<para>
159-
Fixing this is on the to-do list.
160-
</para>
161-
</note>
162-
163163
<para>
164164
The <literal>ARRAY</literal> expression syntax may also be used:
165165
<programlisting>
166166
INSERT INTO sal_emp
167167
VALUES ('Bill',
168168
ARRAY[10000, 10000, 10000, 10000],
169-
ARRAY[['meeting', 'lunch'], ['','']]);
169+
ARRAY[['meeting', 'lunch'], ['training', 'presentation']]);
170170

171171
INSERT INTO sal_emp
172172
VALUES ('Carol',
173173
ARRAY[20000, 25000, 25000, 25000],
174-
ARRAY[['talk', 'consult'], ['meeting', '']]);
175-
SELECT * FROM sal_emp;
176-
name | pay_by_quarter | schedule
177-
-------+---------------------------+-------------------------------
178-
Bill | {10000,10000,10000,10000} | {{meeting,lunch},{"",""}}
179-
Carol | {20000,25000,25000,25000} | {{talk,consult},{meeting,""}}
180-
(2 rows)
174+
ARRAY[['breakfast', 'consulting'], ['meeting', 'lunch']]);
181175
</programlisting>
182-
Note that with this syntax, multidimensional arrays must have matching
183-
extents for each dimension. A mismatch causes an error report, rather than
184-
silently discarding values as in the previous case.
185-
For example:
186-
<programlisting>
187-
INSERT INTO sal_emp
188-
VALUES ('Carol',
189-
ARRAY[20000, 25000, 25000, 25000],
190-
ARRAY[['talk', 'consult'], ['meeting']]);
191-
ERROR: multidimensional arrays must have array expressions with matching dimensions
192-
</programlisting>
193-
Also notice that the array elements are ordinary SQL constants or
176+
Notice that the array elements are ordinary SQL constants or
194177
expressions; for instance, string literals are single quoted, instead of
195178
double quoted as they would be in an array literal. The <literal>ARRAY</>
196179
expression syntax is discussed in more detail in <xref
@@ -247,9 +230,9 @@ SELECT pay_by_quarter[3] FROM sal_emp;
247230
<programlisting>
248231
SELECT schedule[1:2][1:1] FROM sal_emp WHERE name = 'Bill';
249232

250-
schedule
251-
--------------------
252-
{{meeting},{""}}
233+
schedule
234+
------------------------
235+
{{meeting},{training}}
253236
(1 row)
254237
</programlisting>
255238

@@ -266,9 +249,10 @@ SELECT schedule[1:2][1] FROM sal_emp WHERE name = 'Bill';
266249
is specified, as in this example:
267250
<programlisting>
268251
SELECT schedule[1:2][2] FROM sal_emp WHERE name = 'Bill';
269-
schedule
270-
---------------------------
271-
{{meeting,lunch},{"",""}}
252+
253+
schedule
254+
-------------------------------------------
255+
{{meeting,lunch},{training,presentation}}
272256
(1 row)
273257
</programlisting>
274258
</para>
@@ -546,6 +530,47 @@ SELECT * FROM sal_emp WHERE 10000 = ALL (pay_by_quarter);
546530
up to the next right brace or delimiter is taken as the item value.
547531
</para>
548532

533+
<para>
534+
By default, the lower bound index value of an array's dimensions is
535+
set to one. If any of an array's dimensions has a lower bound index not
536+
equal to one, an additional decoration that indicates the actual
537+
array dimensions will precede the array structure decoration.
538+
The decoration consists of square braces (<literal>[</> and <literal>]</>)
539+
around each array dimension's lower and upper bound indicies, plus
540+
a colon (<literal>:</>) delimiter character inbetween. Delimiting the
541+
array dimension decoration from the array structure decoration is a
542+
single assignment operator (<literal>=</>). For example:
543+
<programlisting>
544+
SELECT 1 || ARRAY[2,3] AS array;
545+
546+
array
547+
---------------
548+
[0:2]={1,2,3}
549+
(1 row)
550+
551+
SELECT ARRAY[1,2] || ARRAY[[3,4]] AS array;
552+
553+
array
554+
--------------------------
555+
[0:1][1:2]={{1,2},{3,4}}
556+
(1 row)
557+
</programlisting>
558+
</para>
559+
560+
<para>
561+
In a similar fashion, an array with non-default indicies may be specified
562+
using the same literal syntax. For example:
563+
<programlisting>
564+
SELECT f1[1][-2][3] AS e1, f1[1][-1][5] AS e2
565+
FROM (SELECT '[1:1][-2:-1][3:5]={{{1,2,3},{4,5,6}}}'::int[] AS f1) AS ss;
566+
567+
e1 | e2
568+
----+----
569+
1 | 6
570+
(1 row)
571+
</programlisting>
572+
</para>
573+
549574
<para>
550575
As shown previously, when writing an array value you may write double
551576
quotes around any individual array

‎src/backend/utils/adt/arrayfuncs.c‎

Lines changed: 88 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -8,7 +8,7 @@
88
*
99
*
1010
* IDENTIFICATION
11-
* $PostgreSQL: pgsql/src/backend/utils/adt/arrayfuncs.c,v 1.105 2004/06/16 01:26:47 tgl Exp $
11+
* $PostgreSQL: pgsql/src/backend/utils/adt/arrayfuncs.c,v 1.106 2004/08/05 03:29:37 joe Exp $
1212
*
1313
*-------------------------------------------------------------------------
1414
*/
@@ -217,7 +217,7 @@ array_in(PG_FUNCTION_ARGS)
217217
errmsg("number of array dimensions (%d) exceeds the maximum allowed (%d)",
218218
ndim, MAXDIM)));
219219

220-
for (q = p; isdigit((unsigned char) *q); q++);
220+
for (q = p; isdigit((unsigned char) *q) || (*q == '-') || (*q == '+'); q++);
221221
if (q == p) /* no digits? */
222222
ereport(ERROR,
223223
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
@@ -229,7 +229,7 @@ array_in(PG_FUNCTION_ARGS)
229229
*q = '\0';
230230
lBound[ndim] = atoi(p);
231231
p = q + 1;
232-
for (q = p; isdigit((unsigned char) *q); q++);
232+
for (q = p; isdigit((unsigned char) *q) || (*q == '-') || (*q == '+'); q++);
233233
if (q == p) /* no digits? */
234234
ereport(ERROR,
235235
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
@@ -270,6 +270,9 @@ array_in(PG_FUNCTION_ARGS)
270270
}
271271
else
272272
{
273+
int ndim_braces,
274+
dim_braces[MAXDIM];
275+
273276
/* If array dimensions are given, expect '=' operator */
274277
if (strncmp(p, ASSGN, strlen(ASSGN)) != 0)
275278
ereport(ERROR,
@@ -278,6 +281,27 @@ array_in(PG_FUNCTION_ARGS)
278281
p += strlen(ASSGN);
279282
while (isspace((unsigned char) *p))
280283
p++;
284+
285+
/*
286+
* intuit dimensions from brace structure -- it better match what
287+
* we were given
288+
*/
289+
if (*p != '{')
290+
ereport(ERROR,
291+
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
292+
errmsg("array value must start with \"{\" or dimension information")));
293+
ndim_braces = ArrayCount(p, dim_braces, typdelim);
294+
if (ndim_braces != ndim)
295+
ereport(ERROR,
296+
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
297+
errmsg("array dimensions incompatible with array literal")));
298+
for (i = 0; i < ndim; ++i)
299+
{
300+
if (dim[i] != dim_braces[i])
301+
ereport(ERROR,
302+
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
303+
errmsg("array dimensions incompatible with array literal")));
304+
}
281305
}
282306

283307
#ifdef ARRAYDEBUG
@@ -303,7 +327,6 @@ array_in(PG_FUNCTION_ARGS)
303327
ereport(ERROR,
304328
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
305329
errmsg("missing left brace")));
306-
307330
dataPtr = ReadArrayStr(p, nitems, ndim, dim, &my_extra->proc, typioparam,
308331
typmod, typdelim, typlen, typbyval, typalign,
309332
&nbytes);
@@ -334,13 +357,18 @@ ArrayCount(char *str, int *dim, char typdelim)
334357
int nest_level = 0,
335358
i;
336359
int ndim = 1,
337-
temp[MAXDIM];
360+
temp[MAXDIM],
361+
nelems[MAXDIM],
362+
nelems_last[MAXDIM];
338363
bool scanning_string = false;
339364
bool eoArray = false;
340365
char *ptr;
341366

342367
for (i = 0; i < MAXDIM; ++i)
368+
{
343369
temp[i] = dim[i] = 0;
370+
nelems_last[i] = nelems[i] = 1;
371+
}
344372

345373
if (strncmp(str, "{}", 2) == 0)
346374
return 0;
@@ -394,6 +422,16 @@ ArrayCount(char *str, int *dim, char typdelim)
394422
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
395423
errmsg("malformed array literal: \"%s\"", str)));
396424
nest_level--;
425+
426+
if ((nelems_last[nest_level] != 1) &&
427+
(nelems[nest_level] != nelems_last[nest_level]))
428+
ereport(ERROR,
429+
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
430+
errmsg("multidimensional arrays must have "
431+
"array expressions with matching "
432+
"dimensions")));
433+
nelems_last[nest_level] = nelems[nest_level];
434+
nelems[nest_level] = 1;
397435
if (nest_level == 0)
398436
eoArray = itemdone = true;
399437
else
@@ -408,7 +446,10 @@ ArrayCount(char *str, int *dim, char typdelim)
408446
break;
409447
default:
410448
if (*ptr == typdelim && !scanning_string)
449+
{
411450
itemdone = true;
451+
nelems[nest_level - 1]++;
452+
}
412453
break;
413454
}
414455
if (!itemdone)
@@ -684,16 +725,24 @@ array_out(PG_FUNCTION_ARGS)
684725
char *p,
685726
*tmp,
686727
*retval,
687-
**values;
688-
bool *needquotes;
728+
**values,
729+
/*
730+
* 33 per dim since we assume 15 digits per number + ':' +'[]'
731+
*
732+
* +2 allows for assignment operator + trailing null
733+
*/
734+
dims_str[(MAXDIM * 33) + 2];
735+
bool *needquotes,
736+
needdims = false;
689737
int nitems,
690738
overall_length,
691739
i,
692740
j,
693741
k,
694742
indx[MAXDIM];
695743
int ndim,
696-
*dim;
744+
*dim,
745+
*lb;
697746
ArrayMetaState *my_extra;
698747

699748
element_type = ARR_ELEMTYPE(v);
@@ -734,6 +783,7 @@ array_out(PG_FUNCTION_ARGS)
734783

735784
ndim = ARR_NDIM(v);
736785
dim = ARR_DIMS(v);
786+
lb = ARR_LBOUND(v);
737787
nitems = ArrayGetNItems(ndim, dim);
738788

739789
if (nitems == 0)
@@ -742,6 +792,19 @@ array_out(PG_FUNCTION_ARGS)
742792
PG_RETURN_CSTRING(retval);
743793
}
744794

795+
/*
796+
* we will need to add explicit dimensions if any dimension
797+
* has a lower bound other than one
798+
*/
799+
for (i = 0; i < ndim; i++)
800+
{
801+
if (lb[i] != 1)
802+
{
803+
needdims = true;
804+
break;
805+
}
806+
}
807+
745808
/*
746809
* Convert all values to string form, count total space needed
747810
* (including any overhead such as escaping backslashes), and detect
@@ -798,12 +861,28 @@ array_out(PG_FUNCTION_ARGS)
798861
*/
799862
for (i = j = 0, k = 1; i < ndim; k *= dim[i++], j += k);
800863

801-
retval = (char *) palloc(overall_length + 2 * j);
864+
/* add explicit dimensions if required */
865+
if (needdims)
866+
{
867+
char *ptr = dims_str;
868+
869+
for (i = 0; i < ndim; i++)
870+
{
871+
sprintf(ptr, "[%d:%d]", lb[i], lb[i] + dim[i] - 1);
872+
ptr += strlen(ptr);
873+
}
874+
*ptr++ = *ASSGN;
875+
*ptr = '\0';
876+
}
877+
878+
retval = (char *) palloc(strlen(dims_str) + overall_length + 2 * j);
802879
p = retval;
803880

804881
#define APPENDSTR(str) (strcpy(p, (str)), p += strlen(p))
805882
#define APPENDCHAR(ch) (*p++ = (ch), *p = '\0')
806883

884+
if (needdims)
885+
APPENDSTR(dims_str);
807886
APPENDCHAR('{');
808887
for (i = 0; i < ndim; indx[i++] = 0);
809888
j = 0;

0 commit comments

Comments
 (0)