From f471aba088bd438a731c598060f11871c7fd1264 Mon Sep 17 00:00:00 2001 From: "okbob@github.com" Date: Wed, 12 Aug 2026 07:03:28 +0200 Subject: [PATCH] introduce 'tam' modifier for date/timestamp formatting MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit glibc returns localized month names in the genitive case, but it also provides alternative month names in the nominative case. The TAM prefix allows you to retrieve this alternative name. example: SELECT to_char(date '2026-08-01', 'TMMONTH'); ┌─────────┐ │ to_char │ ╞═════════╡ │ SRPNA │ └─────────┘ (1 row) SELECT to_char(date '2026-08-01', 'TAMMONTH'); ┌─────────┐ │ to_char │ ╞═════════╡ │ SRPEN │ └─────────┘ (1 row) --- doc/src/sgml/func/func-formatting.sgml | 31 +++- src/backend/utils/adt/formatting.c | 170 +++++++++++++++--- src/backend/utils/adt/pg_locale.c | 16 +- src/include/utils/pg_locale.h | 2 + .../regress/expected/collate.linux.utf8.out | 40 +++++ src/test/regress/sql/collate.linux.utf8.sql | 14 ++ 6 files changed, 241 insertions(+), 32 deletions(-) diff --git a/doc/src/sgml/func/func-formatting.sgml b/doc/src/sgml/func/func-formatting.sgml index e4edaf4f42c..563587dc617 100644 --- a/doc/src/sgml/func/func-formatting.sgml +++ b/doc/src/sgml/func/func-formatting.sgml @@ -477,6 +477,12 @@ ) TMMonth + + TAM prefix + translation alternative mode (use localized alternative month names based on + ; see usage notes) + TAMMonth + SP suffix spell mode (not implemented) @@ -504,8 +510,29 @@ - TM suppresses trailing blanks whether or - not FM is specified. + TM and TAM suppress trailing + blanks whether or not FM is specified. + + + + + + TM and TAM both produce + localized month names according to , + but they can differ for certain languages that inflect month names + (e.g., Slavic languages). In such languages TM + produces the form used together with a day number, which is often the + genitive case, while TAM produces the standalone + (nominative) form. For example, with lc_time set to + a Czech locale, to_char('2026-08-01'::date, 'TMMONTH') + returns SRPNA (genitive), whereas + to_char('2026-08-01'::date, 'TAMMONTH') returns + SRPEN (nominative). For languages that do not draw + this distinction, and for the C locale, the two + modifiers produce the same result. This distinction relies on + alternative month names being available from the underlying operating + system's locale support, so TAM may not be effective + on all platforms. diff --git a/src/backend/utils/adt/formatting.c b/src/backend/utils/adt/formatting.c index effad4c37dd..f5376d0a0a7 100644 --- a/src/backend/utils/adt/formatting.c +++ b/src/backend/utils/adt/formatting.c @@ -553,6 +553,7 @@ do { \ #define DCH_SUFFIX_th 0x04 #define DCH_SUFFIX_SP 0x08 #define DCH_SUFFIX_TM 0x10 +#define DCH_SUFFIX_TAM 0x20 /* * Suffix tests @@ -594,16 +595,25 @@ IS_SUFFIX_TM(uint8 _s) return (_s & DCH_SUFFIX_TM); } +static inline bool +IS_SUFFIX_TAM(uint8 _s) +{ + return (_s & DCH_SUFFIX_TAM); +} + /* * Suffixes definition for DATE-TIME TO/FROM CHAR */ #define TM_SUFFIX_LEN 2 +#define TAM_SUFFIX_LEN 3 static const KeySuffix DCH_suff[] = { {"FM", 2, DCH_SUFFIX_FM, SUFFTYPE_PREFIX}, {"fm", 2, DCH_SUFFIX_FM, SUFFTYPE_PREFIX}, {"TM", TM_SUFFIX_LEN, DCH_SUFFIX_TM, SUFFTYPE_PREFIX}, + {"TAM", TAM_SUFFIX_LEN, DCH_SUFFIX_TAM, SUFFTYPE_PREFIX}, {"tm", 2, DCH_SUFFIX_TM, SUFFTYPE_PREFIX}, + {"tam", 3, DCH_SUFFIX_TAM, SUFFTYPE_PREFIX}, {"TH", 2, DCH_SUFFIX_TH, SUFFTYPE_POSTFIX}, {"th", 2, DCH_SUFFIX_th, SUFFTYPE_POSTFIX}, {"SP", 2, DCH_SUFFIX_SP, SUFFTYPE_POSTFIX}, @@ -2573,6 +2583,50 @@ from_char_seq_search(int *dest, const char **src, const char *const *array, return true; } +/* + * Allow to choose localized names or alternative localized names by + * usage prefix TM or TAM. Simple code, just reduction of repeated code. + */ +static char ** +get_localized_abbrev_months(uint8 suffix, int *suffix_len) +{ + if (IS_SUFFIX_TM(suffix)) + { + *suffix_len = 2; + return localized_abbrev_months; + } + else if (IS_SUFFIX_TAM(suffix)) + { + *suffix_len = 3; + return localized_alt_abbrev_months; + } + else + { + *suffix_len = 0; + return NULL; + } +} + +static char ** +get_localized_full_months(uint8 suffix, int *suffix_len) +{ + if (IS_SUFFIX_TM(suffix)) + { + *suffix_len = 2; + return localized_full_months; + } + else if (IS_SUFFIX_TAM(suffix)) + { + *suffix_len = 3; + return localized_alt_full_months; + } + else + { + *suffix_len = 0; + return NULL; + } +} + /* * Process a TmToChar struct as denoted by a list of FormatNodes. * The formatted data is written to the string pointed to by 'out'. @@ -2786,11 +2840,18 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_toupper_z(localized_full_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; - if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) + localized_months = get_localized_full_months(n->suffix, + &suffix_len); + + str = str_toupper_z(localized_months[tm->tm_mon - 1], collid); + + if (strlen(str) <= (n->key->len + suffix_len) * DCH_MAX_ITEM_SIZ) strcpy(s, str); else ereport(ERROR, @@ -2806,11 +2867,18 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_initcap_z(localized_full_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; - if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) + localized_months = get_localized_full_months(n->suffix, + &suffix_len); + + str = str_initcap_z(localized_months[tm->tm_mon - 1], collid); + + if (strlen(str) <= (n->key->len + suffix_len) * DCH_MAX_ITEM_SIZ) strcpy(s, str); else ereport(ERROR, @@ -2826,11 +2894,18 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_tolower_z(localized_full_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; - if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) + localized_months = get_localized_full_months(n->suffix, + &suffix_len); + + str = str_tolower_z(localized_months[tm->tm_mon - 1], collid); + + if (strlen(str) <= (n->key->len + suffix_len) * DCH_MAX_ITEM_SIZ) strcpy(s, str); else ereport(ERROR, @@ -2846,9 +2921,16 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_toupper_z(localized_abbrev_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; + + localized_months = get_localized_abbrev_months(n->suffix, + &suffix_len); + + str = str_toupper_z(localized_months[tm->tm_mon - 1], collid); if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) strcpy(s, str); @@ -2865,9 +2947,16 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_initcap_z(localized_abbrev_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; + + localized_months = get_localized_abbrev_months(n->suffix, + &suffix_len); + + str = str_initcap_z(localized_months[tm->tm_mon - 1], collid); if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) strcpy(s, str); @@ -2884,9 +2973,16 @@ DCH_to_char(FormatNode *node, bool is_interval, TmToChar *in, char *out, Oid col INVALID_FOR_INTERVAL; if (!tm->tm_mon) break; - if (IS_SUFFIX_TM(n->suffix)) + if (IS_SUFFIX_TM(n->suffix) || IS_SUFFIX_TAM(n->suffix)) { - char *str = str_tolower_z(localized_abbrev_months[tm->tm_mon - 1], collid); + char *str; + char **localized_months; + int suffix_len; + + localized_months = get_localized_abbrev_months(n->suffix, + &suffix_len); + + str = str_tolower_z(localized_months[tm->tm_mon - 1], collid); if (strlen(str) <= (n->key->len + TM_SUFFIX_LEN) * DCH_MAX_ITEM_SIZ) strcpy(s, str); @@ -3565,24 +3661,40 @@ DCH_from_char(FormatNode *node, const char *in, TmFromChar *out, case DCH_MONTH: case DCH_Month: case DCH_month: - if (!from_char_seq_search(&value, &s, months_full, - IS_SUFFIX_TM(n->suffix) ? localized_full_months : NULL, - collid, - n, escontext)) - return; - if (!from_char_set_int(&out->mm, value + 1, n, escontext)) - return; + { + char **localized_months; + int suffix_len; + + localized_months = get_localized_full_months(n->suffix, + &suffix_len); + + if (!from_char_seq_search(&value, &s, months_full, + localized_months, + collid, + n, escontext)) + return; + if (!from_char_set_int(&out->mm, value + 1, n, escontext)) + return; + } break; case DCH_MON: case DCH_Mon: case DCH_mon: - if (!from_char_seq_search(&value, &s, months, - IS_SUFFIX_TM(n->suffix) ? localized_abbrev_months : NULL, - collid, - n, escontext)) - return; - if (!from_char_set_int(&out->mm, value + 1, n, escontext)) - return; + { + char **localized_months; + int suffix_len; + + localized_months = get_localized_abbrev_months(n->suffix, + &suffix_len); + + if (!from_char_seq_search(&value, &s, months, + localized_months, + collid, + n, escontext)) + return; + if (!from_char_set_int(&out->mm, value + 1, n, escontext)) + return; + } break; case DCH_MM: if (from_char_parse_int(&out->mm, &s, n, escontext) < 0) diff --git a/src/backend/utils/adt/pg_locale.c b/src/backend/utils/adt/pg_locale.c index 9eb99487e57..773a563867e 100644 --- a/src/backend/utils/adt/pg_locale.c +++ b/src/backend/utils/adt/pg_locale.c @@ -101,7 +101,9 @@ int icu_validation_level = WARNING; char *localized_abbrev_days[7 + 1]; char *localized_full_days[7 + 1]; char *localized_abbrev_months[12 + 1]; +char *localized_alt_abbrev_months[12 + 1]; char *localized_full_months[12 + 1]; +char *localized_alt_full_months[12 + 1]; static pg_locale_t default_locale = NULL; @@ -701,7 +703,7 @@ cache_single_string(char **dst, const char *src, int encoding) void cache_locale_time(void) { - char buf[(2 * 7 + 2 * 12) * MAX_L10N_DATA]; + char buf[(2 * 7 + 4 * 12) * MAX_L10N_DATA]; char *bufptr; time_t timenow; struct tm *timeinfo; @@ -765,9 +767,15 @@ cache_locale_time(void) if (strftime_l(bufptr, MAX_L10N_DATA, "%b", timeinfo, locale) <= 0) strftimefail = true; bufptr += MAX_L10N_DATA; + if (strftime_l(bufptr, MAX_L10N_DATA, "%Ob", timeinfo, locale) <= 0) + strftimefail = true; + bufptr += MAX_L10N_DATA; if (strftime_l(bufptr, MAX_L10N_DATA, "%B", timeinfo, locale) <= 0) strftimefail = true; bufptr += MAX_L10N_DATA; + if (strftime_l(bufptr, MAX_L10N_DATA, "%OB", timeinfo, locale) <= 0) + strftimefail = true; + bufptr += MAX_L10N_DATA; } #ifdef WIN32 @@ -822,11 +830,17 @@ cache_locale_time(void) { cache_single_string(&localized_abbrev_months[i], bufptr, encoding); bufptr += MAX_L10N_DATA; + cache_single_string(&localized_alt_abbrev_months[i], bufptr, encoding); + bufptr += MAX_L10N_DATA; cache_single_string(&localized_full_months[i], bufptr, encoding); bufptr += MAX_L10N_DATA; + cache_single_string(&localized_alt_full_months[i], bufptr, encoding); + bufptr += MAX_L10N_DATA; } localized_abbrev_months[12] = NULL; + localized_alt_abbrev_months[12] = NULL; localized_full_months[12] = NULL; + localized_alt_full_months[12] = NULL; CurrentLCTimeValid = true; } diff --git a/src/include/utils/pg_locale.h b/src/include/utils/pg_locale.h index b74821fdfa9..a23876b5f43 100644 --- a/src/include/utils/pg_locale.h +++ b/src/include/utils/pg_locale.h @@ -42,7 +42,9 @@ extern PGDLLIMPORT int icu_validation_level; extern PGDLLIMPORT char *localized_abbrev_days[]; extern PGDLLIMPORT char *localized_full_days[]; extern PGDLLIMPORT char *localized_abbrev_months[]; +extern PGDLLIMPORT char *localized_alt_abbrev_months[]; extern PGDLLIMPORT char *localized_full_months[]; +extern PGDLLIMPORT char *localized_alt_full_months[]; extern bool check_locale(int category, const char *locale, char **canonname); extern char *pg_perm_setlocale(int category, const char *locale); diff --git a/src/test/regress/expected/collate.linux.utf8.out b/src/test/regress/expected/collate.linux.utf8.out index e0a39e4c300..f4b0d60ea4c 100644 --- a/src/test/regress/expected/collate.linux.utf8.out +++ b/src/test/regress/expected/collate.linux.utf8.out @@ -463,7 +463,28 @@ SELECT to_char(date '2010-04-01', 'DD TMMON YYYY' COLLATE "tr_TR"); 01 NİS 2010 (1 row) +-- to_char +SET lc_time TO 'cs_CZ'; +SELECT to_char(date '2010-02-01', 'DD TMMONTH'); + to_char +---------- + 01 ÚNORA +(1 row) + +SELECT to_char(date '2010-02-01', 'TAMMONTH'); + to_char +--------- + ÚNOR +(1 row) + +SELECT to_char(date '2010-02-01', 'TAMMON'); + to_char +--------- + ÚNO +(1 row) + -- to_date +SET lc_time TO 'tr_TR'; SELECT to_date('01 ŞUB 2010', 'DD TMMON YYYY'); to_date ------------ @@ -497,6 +518,25 @@ SELECT to_date('2010 01 araLık', 'YYYY DD TMMONTH'); 12-01-2010 (1 row) +SET lc_time TO 'cs_CZ'; +SELECT to_date('1 srpna 2010', 'DD TMMONTH YYYY'); + to_date +------------ + 08-01-2010 +(1 row) + +SELECT to_date('1 srp 2010', 'DD TAMMON YYYY'); + to_date +------------ + 08-01-2010 +(1 row) + +SELECT to_date('1 srpen 2010', 'DD TAMMONTH YYYY'); + to_date +------------ + 08-01-2010 +(1 row) + -- backwards parsing CREATE VIEW collview1 AS SELECT * FROM collate_test1 WHERE b COLLATE "C" >= 'bbc'; CREATE VIEW collview2 AS SELECT a, b FROM collate_test1 ORDER BY b COLLATE "C"; diff --git a/src/test/regress/sql/collate.linux.utf8.sql b/src/test/regress/sql/collate.linux.utf8.sql index 6d726ee9c99..544679fede5 100644 --- a/src/test/regress/sql/collate.linux.utf8.sql +++ b/src/test/regress/sql/collate.linux.utf8.sql @@ -182,7 +182,15 @@ SELECT to_char(date '2010-02-01', 'DD TMMON YYYY' COLLATE "tr_TR"); SELECT to_char(date '2010-04-01', 'DD TMMON YYYY'); SELECT to_char(date '2010-04-01', 'DD TMMON YYYY' COLLATE "tr_TR"); +-- to_char +SET lc_time TO 'cs_CZ'; + +SELECT to_char(date '2010-02-01', 'DD TMMONTH'); +SELECT to_char(date '2010-02-01', 'TAMMONTH'); +SELECT to_char(date '2010-02-01', 'TAMMON'); + -- to_date +SET lc_time TO 'tr_TR'; SELECT to_date('01 ŞUB 2010', 'DD TMMON YYYY'); SELECT to_date('01 Şub 2010', 'DD TMMON YYYY'); @@ -192,6 +200,12 @@ SELECT to_date('01 Aralık 2010', 'DD TMMONTH YYYY'); SELECT to_date('01 aralık 2010', 'DD TMMONTH YYYY'); SELECT to_date('2010 01 araLık', 'YYYY DD TMMONTH'); +SET lc_time TO 'cs_CZ'; + +SELECT to_date('1 srpna 2010', 'DD TMMONTH YYYY'); +SELECT to_date('1 srp 2010', 'DD TAMMON YYYY'); +SELECT to_date('1 srpen 2010', 'DD TAMMONTH YYYY'); + -- backwards parsing CREATE VIEW collview1 AS SELECT * FROM collate_test1 WHERE b COLLATE "C" >= 'bbc'; -- 2.55.0