Fix handling of BC years in to_date/to_timestamp.

Previously, a conversion such as
	to_date('-44-02-01','YYYY-MM-DD')
would result in '0045-02-01 BC', as the code attempted to interpret
the negative year as BC, but failed to apply the correction needed
for our internal handling of BC years.  Fix the off-by-one problem.

Also, arrange for the combination of a negative year and an
explicit "BC" marker to cancel out and produce AD.  This is how
the negative-century case works, so it seems sane to do likewise.

Continue to read "year 0000" as 1 BC.  Oracle would throw an error,
but we've accepted that case for a long time so I'm hesitant to
change it in a back-patch.

Per bug #16419 from Saeed Hubaishan.  Back-patch to all supported
branches.

Dar Alathar-Yemen and Tom Lane

Discussion: https://postgr.es/m/16419-d8d9db0a7553f01b@postgresql.org

diff --git a/doc/src/sgml/func.sgml b/doc/src/sgml/func.sgml
index e049a24dfd5369277426e826d5931866149853c1..4313dc5a1e54b2153a603d6a2b3f94581919b822 100644 (file)
--- a/doc/src/sgml/func.sgml
+++ b/doc/src/sgml/func.sgml
@@ -6342,6 +6342,15 @@ SELECT regexp_match('abc01234xyz', '(?:(.*?)(\d+)(.*)){1,1}');
       </para>
      </listitem>
 
+     <listitem>
+      <para>
+       In <function>to_timestamp</function> and <function>to_date</function>,
+       negative years are treated as signifying BC.  If you write both a
+       negative year and an explicit <literal>BC</literal> field, you get AD
+       again.  An input of year zero is treated as 1 BC.
+      </para>
+     </listitem>
+
      <listitem>
       <para>
        In <function>to_timestamp</function> and <function>to_date</function>,

diff --git a/src/backend/utils/adt/formatting.c b/src/backend/utils/adt/formatting.c
index 0bcfca306a6872ef84b31c91ed100c052baedf42..79fccbab76c3b39b20b151366b5b06358c01667f 100644 (file)
--- a/src/backend/utils/adt/formatting.c
+++ b/src/backend/utils/adt/formatting.c
@@ -3754,8 +3754,11 @@ do_to_timestamp(text *date_txt, text *fmt,
                {
                        /* If a 4-digit year is provided, we use that and ignore CC. */
                        tm->tm_year = tmfc.year;
-                       if (tmfc.bc && tm->tm_year > 0)
-                               tm->tm_year = -(tm->tm_year - 1);
+                       if (tmfc.bc)
+                               tm->tm_year = -tm->tm_year;
+                       /* correct for our representation of BC years */
+                       if (tm->tm_year < 0)
+                               tm->tm_year++;
                }
                fmask |= DTK_M(YEAR);
        }

diff --git a/src/test/regress/expected/horology.out b/src/test/regress/expected/horology.out
index 5c4e2a6fcf6c21dddb9e53dec21a06ac281bde35..c3e4b41b3638006b573121183d91bf2df00a9ee1 100644 (file)
--- a/src/test/regress/expected/horology.out
+++ b/src/test/regress/expected/horology.out
@@ -3011,6 +3011,45 @@ SELECT to_date('2458872', 'J');
  01-23-2020
 (1 row)
 
+--
+-- Check handling of BC dates
+--
+SELECT to_date('44-02-01 BC','YYYY-MM-DD BC');
+    to_date    
+---------------
+ 02-01-0044 BC
+(1 row)
+
+SELECT to_date('-44-02-01','YYYY-MM-DD');
+    to_date    
+---------------
+ 02-01-0044 BC
+(1 row)
+
+SELECT to_date('-44-02-01 BC','YYYY-MM-DD BC');
+  to_date   
+------------
+ 02-01-0044
+(1 row)
+
+SELECT to_timestamp('44-02-01 11:12:13 BC','YYYY-MM-DD HH24:MI:SS BC');
+          to_timestamp           
+---------------------------------
+ Fri Feb 01 11:12:13 0044 PST BC
+(1 row)
+
+SELECT to_timestamp('-44-02-01 11:12:13','YYYY-MM-DD HH24:MI:SS');
+          to_timestamp           
+---------------------------------
+ Fri Feb 01 11:12:13 0044 PST BC
+(1 row)
+
+SELECT to_timestamp('-44-02-01 11:12:13 BC','YYYY-MM-DD HH24:MI:SS BC');
+         to_timestamp         
+------------------------------
+ Mon Feb 01 11:12:13 0044 PST
+(1 row)
+
 --
 -- Check handling of multiple spaces in format and/or input
 --
@@ -3198,6 +3237,12 @@ SELECT to_date('2016 366', 'YYYY DDD');  -- ok
 
 SELECT to_date('2016 367', 'YYYY DDD');
 ERROR:  date/time field value out of range: "2016 367"
+SELECT to_date('0000-02-01','YYYY-MM-DD');  -- allowed, though it shouldn't be
+    to_date    
+---------------
+ 02-01-0001 BC
+(1 row)
+
 --
 -- Check behavior with SQL-style fixed-GMT-offset time zone (cf bug #8572)
 --

diff --git a/src/test/regress/sql/horology.sql b/src/test/regress/sql/horology.sql
index 9ac3f2344844ddf10c8cf69619316b3f419b77d6..e479e335172fa62ddaa233075aeda0c9677bfd1d 100644 (file)
--- a/src/test/regress/sql/horology.sql
+++ b/src/test/regress/sql/horology.sql
@@ -467,6 +467,17 @@ SELECT to_date('1 4 1902', 'Q MM YYYY');  -- Q is ignored
 SELECT to_date('3 4 21 01', 'W MM CC YY');
 SELECT to_date('2458872', 'J');
 
+--
+-- Check handling of BC dates
+--
+
+SELECT to_date('44-02-01 BC','YYYY-MM-DD BC');
+SELECT to_date('-44-02-01','YYYY-MM-DD');
+SELECT to_date('-44-02-01 BC','YYYY-MM-DD BC');
+SELECT to_timestamp('44-02-01 11:12:13 BC','YYYY-MM-DD HH24:MI:SS BC');
+SELECT to_timestamp('-44-02-01 11:12:13','YYYY-MM-DD HH24:MI:SS');
+SELECT to_timestamp('-44-02-01 11:12:13 BC','YYYY-MM-DD HH24:MI:SS BC');
+
 --
 -- Check handling of multiple spaces in format and/or input
 --
@@ -535,6 +546,7 @@ SELECT to_date('2015 366', 'YYYY DDD');
 SELECT to_date('2016 365', 'YYYY DDD');  -- ok
 SELECT to_date('2016 366', 'YYYY DDD');  -- ok
 SELECT to_date('2016 367', 'YYYY DDD');
+SELECT to_date('0000-02-01','YYYY-MM-DD');  -- allowed, though it shouldn't be
 
 --
 -- Check behavior with SQL-style fixed-GMT-offset time zone (cf bug #8572)