Casting a DATE literal without time part to TIME datatype fails

contrib/babelfishpg_tsql/src/hooks.c

@@ -41,10 +41,13 @@
 #include "parser/parser.h"
 #include "parser/scanner.h"
 #include "parser/scansup.h"
+#include "regex/regex.h"
 #include "replication/logical.h"
 #include "rewrite/rewriteHandler.h"
 #include "tcop/utility.h"
 #include "utils/builtins.h"
+#include "utils/date.h"
+#include "utils/datetime.h"
 #include "utils/fmgroids.h"
 #include "utils/lsyscache.h"
 #include "utils/rel.h"
@@ -139,6 +142,9 @@ static bool pltsql_bbfCustomProcessUtility(ParseState *pstate,
 									  const char *queryString,
 									  ProcessUtilityContext context,
 									  ParamListInfo params, QueryCompletion *qc);
+static bool isTextMonthPresent(char* field);
+static bool containsInTextMonthFormat(int *ftype, char **field);
+static Datum pltsql_time_in(const char *str, int32 typmod);
 static void pltsql_bbfSelectIntoAddIdentity(IntoClause *into,  List *tableElts);
 extern void pltsql_bbfSelectIntoUtility(ParseState *pstate, PlannedStmt *pstmt, const char *queryString, 
 					QueryEnvironment *queryEnv, ParamListInfo params, QueryCompletion *qc);
@@ -224,6 +230,7 @@ static table_variable_satisfies_update_hook_type prev_table_variable_satisfies_u
 static table_variable_satisfies_vacuum_hook_type prev_table_variable_satisfies_vacuum = NULL;
 static table_variable_satisfies_vacuum_horizon_hook_type prev_table_variable_satisfies_vacuum_horizon = NULL;
 static drop_relation_refcnt_hook_type prev_drop_relation_refcnt_hook = NULL;
+static tsql_time_in_hook_type prev_tsql_time_in_hook = NULL;
 
 /*****************************************
  * 			Install / Uninstall
@@ -384,6 +391,9 @@ InstallExtendedHooks(void)
 
 	prev_drop_relation_refcnt_hook = drop_relation_refcnt_hook;
 	drop_relation_refcnt_hook = pltsql_drop_relation_refcnt_hook;
+
+	prev_tsql_time_in_hook = tsql_time_in_hook;
+	tsql_time_in_hook = pltsql_time_in;
 }
 
 void
@@ -446,6 +456,7 @@ UninstallExtendedHooks(void)
 	IsToastRelationHook = PrevIsToastRelationHook;
 	IsToastClassHook = PrevIsToastClassHook;
 	drop_relation_refcnt_hook = prev_drop_relation_refcnt_hook;
+	tsql_time_in_hook = prev_tsql_time_in_hook;
 }
 
 /*****************************************
@@ -2822,6 +2833,255 @@ pltsql_detect_numeric_overflow(int weight, int dscale, int first_block, int nume
 	return (total_digit_count > TDS_NUMERIC_MAX_PRECISION);
 }
 
+/* Checks whether the field is valid text month */
+static bool isTextMonthPresent(char* field)
+{
+	char* months[] = {"january", "february", "march", "april", "may",
+					"june", "july", "august", "september", "october",
+					"november", "december", "jan", "feb", "mar", "apr",
+					"may", "jun", "jul", "aug", "sep", "oct", "nov", "dec"};
+	for(int i = 0; i < 24; i++)
+		if(pg_strcasecmp(field, months[i]) == 0)
+			return true;
+	return false;
+}
+
+/*
+ * This function will check whether the first 3 inputs are in any of the format
+ * `DD MON YYY`, `DD YYYY MON`, `MON DD YYYY`, `MON YYYY DD`, `YYYY MM DD`, `YYYY DD MM`
+ * where MON is month in text format then returns true if present.
+ */
+static bool containsInTextMonthFormat(int *ftype, char **field){
+	int count_number = 0, count_string = 0;
+	for(int i = 0; i < 3; i++)
+	{
+		if(ftype[i] == DTK_NUMBER)
+			count_number++;
+		else if(ftype[i] == DTK_STRING)
+		{
+			/* Check whether the string is valid text month */
+			if(!isTextMonthPresent(field[i]))
+				return false;
+			count_string++;
+		}
+		else
+			return false;
+	}
+	if(count_number == 2 && count_string == 1)
+	{
+		/*
+		 * If the first field is an string then swap with the second field
+		 * as when the date is given separatly then all different forms of
+		 * dates is supported. To avoid the conversion failure from `isTextMonthPresent`
+		 * later we are swapping earlier.
+		 */
+		if(ftype[0] == DTK_STRING)
+		{
+			char* temp_field;
+			int temp_ftype;
+			temp_field = field[0];
+			temp_ftype = ftype[0];
+			field[0] = field[1];
+			field[1] = temp_field;
+			ftype[0] = ftype[1];
+			ftype[1] = temp_ftype;
+		}
+		return true;
+	}
+	return false;
+}
+
+Datum pltsql_time_in(const char *str, int32	typmod)
+{
+	TimeADT		result;
+	fsec_t		fsec;
+	struct pg_tm tt,
+			   *tm = &tt;
+	int			tz;
+	int			nf;
+	int			dterr;
+	char		workbuf[MAXDATELEN + 1];
+	char	   *field[MAXDATEFIELDS];
+	int			dtype;
+	int			ftype[MAXDATEFIELDS];
+	StringInfo	res ;
+	int status;
+
+	/* Throw a common error message while casting to time datatype */
+	#define TIME_IN_ERROR()	\
+		ereport(ERROR,	\
+			(errcode(ERRCODE_INVALID_DATETIME_FORMAT),	\
+			errmsg("Conversion failed when converting date and/or time from character string.")));	\
+
+	dterr = ParseDateTime(str, workbuf, sizeof(workbuf),
+						field, ftype, MAXDATEFIELDS, &nf);
+	if (dterr == 0)
+	{
+		if(nf >= 3)
+		{
+			/*
+			 * If the input is of format "YYYY MON DD", then convert
+			 * to format of YYYY-MON-DD and ftype changes to "DTK_DATE"
+			 */
+			if(containsInTextMonthFormat(ftype, field))
+			{
+				/*
+				 * For example if input is "2000 nov 23" will be converted
+				 * to "2000-nov-23".
+				 */
+				res = makeStringInfo();
+				appendStringInfo(res, "%s-%s-%s", field[0], field[1], field[2]);
+				field[0] = NULL;
+				field[0] = res->data;
+				ftype[0] = DTK_DATE;
+				nf = nf - 2;
+				/*
+				 * Since the first 3 fields converted to 1,
+				 * skip the attached fields.
+				 */
+				for(int i = 1; i < nf; i++)
+				{
+					field[i] = field[i+2];
+					ftype[i] = ftype[i+2];
+				}
+			}
+		}
+
+		for(int i = 0; i < nf; i++)
+		{
+			char *temp_field;
+			int len;
+			char* different_date_formats[] = {"/", ".", "-"};
+			temp_field = pstrdup(field[i]);
+			len = strlen(temp_field);
+			switch(ftype[i])
+			{
+				case DTK_NUMBER:
+					switch(len)
+					{
+						case 1:
+						case 2:
+							/*
+							 * At this point the digit number of length 1,2 should be considered as
+							 * time only if there is an [ap]m after the number or else
+							 * should throw an error
+							 */
+							if(i == nf - 1 || ftype[i+1] != DTK_STRING)
+								TIME_IN_ERROR();
+							status = RE_compile_and_execute(cstring_to_text("^([ap]m)$"),
+															(char *) field[i+1], strlen(field[i+1]),
+															REG_ADVANCED, DEFAULT_COLLATION_OID,
+															0, NULL);
+							if(!status)
+								TIME_IN_ERROR();
+							/*
+							 * For example if the input is "1 am"
+							 * will be converted to "1:00:00 am".
+							 */
+							res = makeStringInfo();
+							appendStringInfo(res, "%s%s", field[i], ":00:00");
+							field[i] = NULL;
+							field[i] = res->data;
+							ftype[i] = DTK_TIME;
+							break;
+
+						case 4:
+							/*
+							 * If the numeric input is of length 4, then convert
+							 * to year with default format of YYYY-01-01.
+							 * For example if input is "2000" will be converted
+							 * to "2000-01-01".
+							 */
+							res = makeStringInfo();
+							appendStringInfo(res, "%s%s", field[i], "-01-01");
+							field[i] = NULL;
+							field[i] = res->data;
+							ftype[i] = DTK_DATE;
+							break;
+
+						case 6:
+						case 8:
+							res = makeStringInfo();
+							for(int k = 0; k < len; k++)
+							{
+								appendStringInfo(res, "%c", temp_field[k]);
+								if((len == 6 && ((k + 1) % 2 == 0 && k < 5)) ||
+									(len == 8 && (k == 3 || k == 5)))
+										appendStringInfo(res, "%c", '-');
+							}
+							field[i] = NULL;
+							field[i] = res->data;	
+							ftype[i] = DTK_DATE;
+							break;
+						/*
+						 * If the numeric is of length {3,5,7, >8} then an error should be thrown*/
+						default:
+							TIME_IN_ERROR();
+					}
+					break;
+
+				case DTK_DATE:
+					/*
+					 * If the input is of format `Mon{/.-}yyyy{/.-}dd` or `Mon{/.-}dd{/.-}yyyy
+					 * shouldn't be supported.
+					 * Supported date formats are `Mon yyyy dd`, `Mon dd yyyy`, `mm{-/.}dd{-/.}yyyy`, `yyyy{-/.}mm{-/.}dd`
+					 */
+					for(int k = 0 ; k < 3; k++)
+					{
+						temp_field = strtok(temp_field, different_date_formats[k]);
+						if (pg_strcasecmp(field[i], temp_field) != 0)
+							break;
+					}
+					if(isTextMonthPresent(temp_field))
+						TIME_IN_ERROR();
+					break;
+
+				default:
+					break;
+			}
+			pfree(temp_field);
+		}
+		switch(nf)
+		{
+			case 1:
+				/*
+				 * If only date is specified add an default time of
+				 * 00:00:00
+				 */
+				if(ftype[0] == DTK_DATE)
+				{
+					ftype[1] = DTK_TIME;
+					field[1] = "00:00:00";
+					nf = nf + 1;
+ 				}
+				break;
+			case 3:
+				/*
+				 * If the given input is of format `yyyy-mm-ddThh:mm:ss`
+				 * then conver it to `yyyy-mm-dd hh:mm:ss` by ignoring 'T'
+				 */
+				if(ftype[1] == DTK_STRING && pg_strcasecmp(field[1], "t") == 0 && ftype[2] == DTK_TIME)
+				{
+					ftype[1] = ftype[2];
+					field[1] = field[2];
+					nf = nf - 1;
+				}
+			default:
+				break;
+		}
+
+		dterr = DecodeTimeOnly(field, ftype, nf, &dtype, tm, &fsec, &tz);
+	}
+
+	if (dterr != 0)
+		TIME_IN_ERROR();
+
+	tm2time(tm, fsec, &result);
+	AdjustTimeForTypmod(&result, typmod);
+
+	PG_RETURN_TIMEADT(result);
+}
+
 /*
  * Stores argument positions of default values of a PL/tsql function to bbf_function_ext catalog
  */

contrib/babelfishpg_tsql/src/tsqlIface.cpp

@@ -3,6 +3,7 @@
 #include <iostream>
 #include <strstream>
 #include <string>
+#include <regex>
 #include <unordered_map>
 
 #pragma GCC diagnostic ignored "-Wattributes"
@@ -1990,6 +1991,37 @@ class tsqlBuilder : public tsqlCommonMutator
 			/* Re-write session_user to sys.session_user(). */
 			if (bctx->bif_no_brackets && bctx->SESSION_USER())
 				rewritten_query_fragment.emplace(std::make_pair(bctx->bif_no_brackets->getStartIndex(), std::make_pair(::getFullText(bctx->SESSION_USER()), "sys.session_user()")));
+
+			if(bctx->bif_cast_parse())
+			{
+				auto cast_parse = bctx->bif_cast_parse();
+				/* Re-write the expression for casting to time datatype. */
+				if(cast_parse->data_type() && pg_strncasecmp(::getFullText(cast_parse->data_type()).c_str(), "TIME", 4) == 0)
+				{
+					std::string cast_expression(pstrdup(::getFullText(cast_parse->expression()).c_str()));
+
+					/* 
+					 * Remove spaces between before or after {/:.} in the given expression.
+					 * If the expression contains input of '  dd   /   mm/   yyyy     hh:  mm :   ss  .  ss'
+					 * will be converted to '  dd/mm/yyyy     hh:mm:ss.ss'
+					 */
+					std::regex pattern("\\s*([/:.])\\s*");
+					std::string result_expression = std::regex_replace(cast_expression, pattern, "$1");
+
+					/* 
+					 * Remove spaces in between date in different kind of format dd   {-}   mm   {-}   yyyy.
+					 * The {-} was not included above as the timezone can contain {+,-} which shouldn't be affected.
+					 * For example if the expression contains ' dd - mm -yyyy  hh:mm'
+					 * will be converted to ' dd-mm-yyyy hh:mm'
+					 */
+					pattern = ("\\s*(\\d{2,4}|\\w+)\\s*(-)\\s*(\\d{1,2}|\\w+)\\s*(-)\\s*(\\d{1,4}|\\w+)\\s*");
+					result_expression = std::regex_replace(result_expression, pattern, "$1$2$3$4$5 ");
+
+					/* If the expression didn't change after comparing with above regex no need to change */
+					if ( cast_expression!= result_expression)
+						rewritten_query_fragment.emplace(std::make_pair(cast_parse->expression()->start->getStartIndex(), std::make_pair(::getFullText(cast_parse->expression()), result_expression)));
+				}
+			}
 		}
 
 		/* analyze scalar function call */

test/JDBC/expected/babel_time-vu-cleanup.out

@@ -0,0 +1,11 @@
+DROP TABLE babel_time_vu_prepare_t1
+GO
+
+DROP VIEW babel_time_vu_prepare_v1
+GO
+
+DROP PROCEDURE babel_time_vu_prepare_p1
+GO
+
+DROP FUNCTION babel_time_vu_prepare_f1
+GO

test/JDBC/expected/babel_time-vu-prepare.out

@@ -0,0 +1,21 @@
+create table babel_time_vu_prepare_t1(a time)
+GO
+
+insert into babel_time_vu_prepare_t1 values ('2012-02-23')
+GO
+~~ROW COUNT: 1~~
+
+
+create view babel_time_vu_prepare_v1 as select CAST('2012-02-23' AS time) as val
+GO
+
+create procedure babel_time_vu_prepare_p1 as select CAST('2012-02-23' AS time) as val
+GO
+
+create function babel_time_vu_prepare_f1()
+returns time
+as
+begin
+	return (select CAST('2012-02-23' AS time) as val)
+end
+GO