I have the following Coldfusion process:
My code makes a database call to the proc CommentInsert (this inserts a comment, and then calls an event insert proc about the comment being added called EventInsert)
I then call Event.GetEventByCommentId(commentId)
The result is no records returned, as the EventInsert hasn't finished adding the event record triggered by CommentInsert in Step 1.
I know this is the case, because if I create a delay between steps 1 and 2, then a recordset IS returned in step 2.
This leads me to believe that the read in step 2 is happening too quickly, before the event insert has committed in step 1.
My question is, how do tell the Coldfusion process to wait till Step 1 has completed before doing the read in Step 2??
Step one and step two are tow totally separate methods.
Code:
<cfset MessageHandlerManager = AddComment(argumentCollection=arguments) />
<cfset qEvents = application.API.EventManager.GetEventFeed(commentId=MessageHandlerManager.GetReturnItems()) />
Also, just let me add that the commentId being passed is valid. I have checked.
Another way to look at it:
Given this code:
<!--- Calls CommentInsert proc, which inserts a comment AND inserts an
event record by calling EventInsert within the proc --->
<cfset var newCommentId = AddComment(argumentCollection=arguments) />
<cfloop from="1" to="1000000" index="i">
</cfloop>
<!--- Gets the event record inserted in the code above --->
<cfset qEvent =
application.API.EventManager.GetEventFeed(commentId=newCommentId ) />
When I run the above code, qEvent comes back with a valid record.
However, when I comment out the loop, the record is coming back
empty.
What I think is happening is that the CommentInsert returns the new
comment Id, but when the GetEventFeed function is called, the
EventInsert proc hasn't completed in time and no record is found.
Thus, by adding the loop and delaying a bit, the event insert has time
to finish and then a valid record is returned when GetEventFeed is
called.
So my question is, how do I prevent this without using the loop.
UPDATE:
Here are the two stored procs used:
DELIMITER $$
DROP PROCEDURE IF EXISTS `CommentInsert` $$
CREATE DEFINER=`root`#`%` PROCEDURE `CommentInsert`(
IN _commentParentId bigint,
IN _commentObjectType int,
IN _commentObjectId bigint,
IN _commentText text,
IN _commentAuthorName varchar(100),
IN _commentAuthorEmail varchar(255),
IN _commentAuthorWebsite varchar(512),
IN _commentSubscribe tinyint(1),
IN _commentIsDisabled tinyint(1),
IN _commentIsActive tinyint(1),
IN _commentCSI int,
IN _commentCSD datetime,
IN _commentUSI int,
IN _commentUSD datetime,
OUT _commentIdOut bigint
)
BEGIN
DECLARE _commentId bigint default 0;
INSERT INTO comment
(
commentParentId,
commentObjectType,
commentObjectId,
commentText,
commentAuthorName,
commentAuthorEmail,
commentAuthorWebsite,
commentSubscribe,
commentIsDisabled,
commentIsActive,
commentCSI,
commentCSD,
commentUSI,
commentUSD
)
VALUES
(
_commentParentId,
_commentObjectType,
_commentObjectId,
_commentText,
_commentAuthorName,
_commentAuthorEmail,
_commentAuthorWebsite,
_commentSubscribe,
_commentIsDisabled,
_commentIsActive,
_commentCSI,
_commentCSD,
_commentUSI,
_commentUSD
);
SET _commentId = LAST_INSERT_ID();
CALL EventInsert(6, Now(), _commentId, _commentObjectType, _commentObjectId, null, null, 'Comment Added', 1, _commentCSI, Now(), _commentUSI, Now());
SELECT _commentId INTO _commentIdOut ;
END $$
DELIMITER ;
DELIMITER $$
DROP PROCEDURE IF EXISTS `EventInsert` $$
CREATE DEFINER=`root`#`%` PROCEDURE `EventInsert`(
IN _eventTypeId int,
IN _eventCreateDate datetime,
IN _eventObjectId bigint,
IN _eventAffectedObjectType1 int,
IN _eventAffectedObjectId1 bigint,
IN _eventAffectedObjectType2 int,
IN _eventAffectedObjectId2 bigint,
IN _eventText varchar(1024),
IN _eventIsActive tinyint,
IN _eventCSI int,
IN _eventCSD datetime,
IN _eventUSI int,
IN _eventUSD datetime
)
BEGIN
INSERT INTO event
(
eventTypeId,
eventCreateDate,
eventObjectId,
eventAffectedObjectType1,
eventAffectedObjectId1,
eventAffectedObjectType2,
eventAffectedObjectId2,
eventText,
eventIsActive,
eventCSI,
eventCSD,
eventUSI,
eventUSD
)
VALUES
(
_eventTypeId,
_eventCreateDate,
_eventObjectId,
_eventAffectedObjectType1,
_eventAffectedObjectId1,
_eventAffectedObjectType2,
_eventAffectedObjectId2,
_eventText,
_eventIsActive,
_eventCSI,
_eventCSD,
_eventUSI,
_eventUSD
);
END $$
DELIMITER ;
Found it. Boiled down to this line in the EventManager.GetEventFeed query:
AND eventCreateDate <= <cfqueryparam cfsqltype="cf_sql_timestamp" value="#Now()#" />
What was happening was the MySql Now() function called in the EventInsert proc was a fraction later than the Coldfusion #Now()# being used in the query. Therefore the line of code excluded that record.Also, why it was only happening when comments were added quickly.
What a biatch. Thanks for the input everyone.
Let me get this straight :
You call a MySQL SP which does an insert which then calls another SP to do another insert.
There's no return to ColdFusion between those two? Is that right?
If that's the case then the chances are there's a problem with your SP not returning values correctly or you're looking in the wrong place for the result.
I'm more inclined towards there being a problem with the MySQL SPs. They aren't exactly great and don't really give you a great deal of performance benefit. Views are useful, but the SPs are, frankly, a bit rubbish. I suspect that when you call the second SP from within the first SP and it returns a value its not being correctly passed back out of the original SP to ColdFusion, hence the lack of result.
To be honest, my suggestion would be to write two ORM functions or simple cfqueries in a suitable DAO or service to record the result of the insert of comment first and return a value. Having returned that value, make the other call to the function to get your event based on the returned comment id. (ColdFusion 8 will give you Generated_Key, ColdFusion 9 is generatedkey, I'm not sure what it'll be in Railo, but it'll be there in the "result" attribute structure).
Thinking about it, I'm not even sure why you're getting the event based on the commentid just entered. You've just added that comment against an event, so you should already have some data on that event, even if its just the ID which you can then get the full event record/object from without having to go around the house via the comment.
So over all I would suggest taking a step back and looking at the data flow you're working with and perhaps refactor it.
Related
I am writing a postgres stored procedure that loops through the rows returned from a select statement. For each row it loops through, it inserts values from that select statement into a new table. One of the values I need to insert into the second table is the average of a column. However, when I call the stored procedure, I get an error that the temporary row has no attribute for the actual column that I am averaging. See stored procedure and error below.
Stored Procedure:
create or replace procedure sendToDataset(sentence int)
as $$
declare temprow peoplereviews%rowtype;
BEGIN
FOR temprow IN
select rulereviewid, avg(rulereview)
from peoplereviews
where sentenceid = sentence
group by rulereviewid
loop
insert into TrainingDataSet(sentenceId, sentence, ruleCorrectId, ruleCorrect, dateAdded)
values(sentence, getSentenceFromID(sentence), tempRow.rulereviewid, tempRow.avg(rulereview), current_timestamp);
END LOOP;
END
$$
LANGUAGE plpgsql;
Error:
ERROR: column "rulereview" does not exist
LINE 2: ...omID(sentence), tempRow.rulereviewid, tempRow.avg(rulereview...
^
QUERY: insert into TrainingDataSet(sentenceId, sentence, ruleCorrectId, ruleCorrect, dateAdded)
values(sentence, getSentenceFromID(sentence), tempRow.rulereviewid, tempRow.avg(rulereview), current_timestamp)
CONTEXT: PL/pgSQL function sendtodataset(integer) line 11 at SQL statement
SQL state: 42703
Basically, I am wondering if it's possible to use that aggregate function in the insert statement or not and if not, if there is another way around it.
you don't need to use a slow and inefficient loop for this:
insert into TrainingDataSet(sentenceId, sentence, ruleCorrectId, ruleCorrect, dateAdded)
select getSentenceId(sentence), sentence, rulereviewid, avg(rulereview), current_timestamp
from peoplereviews
where sentenceid = sentence
group by rulereviewid
To answer the original question: you need to provide a proper alias for the aggregate:
FOR temprow IN
select rulereviewid, avg(rulereview) as avg_views
from peoplereviews
where sentenceid = sentence
group by rulereviewid
loop
insert into TrainingDataSet(sentenceId, sentence, ruleCorrectId, ruleCorrect, dateAdded)
values(sentence, getSentenceFromID(sentence), tempRow.rulereviewid,
tempRow.avg_views, current_timestamp);
END LOOP;
I created sample code using the Microsoft SQL Server Northwind demo database. If you don't have access to this demo database here is a simple (MS-SQL) script to create the table and a row of data for this question.
CREATE TABLE [dbo].[Products](
[ProductID] [int] IDENTITY(1,1) NOT NULL,
[ProductName] [nvarchar](40) NOT NULL,
[SupplierID] [int] NULL,
[CategoryID] [int] NULL,
[QuantityPerUnit] [nvarchar](20) NULL,
[UnitPrice] [money] NULL CONSTRAINT [DF_Products_UnitPrice] DEFAULT (0),
[UnitsInStock] [smallint] NULL CONSTRAINT [DF_Products_UnitsInStock] DEFAULT (0),
[UnitsOnOrder] [smallint] NULL CONSTRAINT [DF_Products_UnitsOnOrder] DEFAULT (0),
[ReorderLevel] [smallint] NULL CONSTRAINT [DF_Products_ReorderLevel] DEFAULT (0),
[Discontinued] [bit] NOT NULL CONSTRAINT [DF_Products_Discontinued] DEFAULT (0),
CONSTRAINT [PK_Products] PRIMARY KEY CLUSTERED
(
[ProductID] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON, FILLFACTOR = 90) ON [PRIMARY]
) ON [PRIMARY]
GO
SET IDENTITY_INSERT [dbo].[Products] ON
GO
INSERT [dbo].[Products] ([ProductID], [ProductName], [SupplierID], [CategoryID], [QuantityPerUnit], [UnitPrice], [UnitsInStock], [UnitsOnOrder], [ReorderLevel], [Discontinued]) VALUES (1, N'Chai', 1, 1, N'10 boxes x 20 bags', 18.0000, 39, 0, 10, 0)
GO
SET IDENTITY_INSERT [dbo].[Products] OFF
GO
Here is the ColdFusion code:
<cfset variables.useTempVar = false>
<cfquery datasource="Northwind2014" name="qryNWProducts">
SELECT TOP 1 * from Products;
</cfquery>
<cfdump var="#qryNWProducts#" label="qryNWProducts">
<cfset variables['stProduct'] = {}>
<cfloop index="vcColName" list="#qryNWProducts.columnlist#">
<cfif variables.useTempVar>
<cfset variables['temp'] = qryNWProducts[vcColName]>
<cfset variables['stProduct'][vcColName] = variables.temp>
<cfelse>
<cfset variables['stProduct'][vcColName] = qryNWProducts[vcColName]>
</cfif>
</cfloop>
<cfdump var="#variables['stProduct']#" label="variables['stProduct']">
<cfloop collection="#variables['stProduct']#" item="key"><cfoutput>
variables['stProduct']['#key#'] JVM datatype = #getMetadata(variables['stProduct'][key]).getName()#<br>
</cfoutput></cfloop>
<br>
This always works:<br>
<cfset variables['aPhrase'] = "I ordered " & variables.stProduct.ProductName & " for " & DollarFormat(variables.stProduct.UnitPrice) & ".">
<cfoutput>#variables['aPhrase']#<br></cfoutput>
<br>
With "variables.useTempVar = false", the next line will throw a "Complex object types cannot be converted to simple values. " error.<br>
<cfset variables['aPhrase'] = "I ordered " & variables['stProduct']['ProductName'] & " for " & DollarFormat(variables['stProduct']['UnitPrice']) & ".">
<cfoutput>#variables['aPhrase']#<br></cfoutput>
The code above has a boolean variable named "variables.useTempVar" at the top that can be flipped to see the error that I'm getting.
It looks like the direct assignment (when variables.useTempVar = false) from the query to the structure causes the structure values to be of JVM type "coldfusion.sql.QueryColumn".
Another note: if this line of code:
<cfset variables['stProduct'][vcColName] = variables.temp>
is changed to:
<cfset variables['stProduct'][vcColName] = variables['temp']>
The JVM datatype will be "coldfusion.sql.QueryColumn".
When the dot notation temp variable is used to assign the query field (when variables.useTempVar = true); the JVM datatypes are simple types that matches up pretty well with the database column types (java.lang.Integer, java.math.BigDecimal, java.lang.String, etc.).
I've also experiemented with statements like this and that provided some odd results:
<cfset variables['stProduct'][vcColName] = qryNWProducts[vcColName].toString()>
Here's the question. Is this the best way to transfer the simple values from a query to a structure? It seems odd to be forced to use a temp variable and dot notation to make this work.
Comment: I've always thought that dot notation and associative array notation were equivalent. This code example appears to contradict that opinion.
#Leigh is correct in that you need to supply the row number when using associative array notation with a query object. So you'd reference row 1 like: qryNWProducts[vcColName][1]
As for your question
Is this the best way to transfer the simple values from a query to a structure?
Are you sure you need a struct? Your question doesn't really specify the use case, so it is entirely possible that you would be better off using the query object as-is.
If you do need it to be a struct (and since you are using ColdFusion 11) might I suggest you take a look at serializeJSON/deSerializeJSON to convert this to a struct. The serializeJSON has a new attribute that will properly serialize a query object into an "AJAX friendly" JSON array of structs. You can then deSerialize the JSON into a CF array, like so:
NWProducts = deSerializeJSON( serializeJSON( qryNWProducts, 'struct' ) )[1];
Which would return a struct representation of the first row in that query object.
Although it's not obvious from the Adobe docs for serializeJSON, the second parameter can be one of: true|false|struct|row|column which will change how the resulting data is formatted.
Here's a runnable example of using the above technique showcasing each serializeQueryAs option.
It's also a better practice to start moving that kind of code into cfscript. queryExecute is quite easy to use and makes script based queries very easy to develop. See the How To Create a Query in cfscript tutorial at trycf.com for more on how to develop script based queries.
Final note, and this is a bit off topic but it is a generally accepted best practice to not use Hungarian Notation when naming variables.
#Abram's covered the mains answer, but just to pick up one tangential point you raise.
Dot notation and associative array notation are generally equivalent in CFML. However in the case of queries, there is a slight variation. Dot notation: query.columnName is treated as shorthand for query.columnName[currentRow] (where currentRow defaults to 1).
Associative array notation with queries does not have this "syntactic sugar", so query["columnName"] refers to the entire column, as the syntax actually indicates.
There are no functions I am aware of in CFML that take a query column as an argument, however the CFML engine will convert the column to an array if it's used in an array function. This is quite handy sometimes.
Can anyone help me to make my code less? (If you can notice to both if-elsif statements I make the same Select.. so I wish there was a way to make this select once. and update with 1 or 0 depending on the pilot_action).
Below its my code.
create or replace
PROCEDURE F_16 (TRK_ID NUMBER, pilot_action NUMBER) IS
BEGIN
BEGIN
IF pilot_action=0 THEN
UPDATE "ControlTow"
SET "Intention"=0
WHERE "Id" IN (
SELECT "Id" FROM "ControlTow" WHERE "Id"=TRK_ID );
ELSIF pilot_action=1 THEN
UPDATE "ControlTow"
SET "Intention"=1
WHERE "Id" IN (
SELECT "Id" FROM "ControlTow" WHERE "Id"=TRK_ID );
END IF;
EXCEPTION
WHEN NO_DATA_FOUND THEN dbms_output.put_line('False Alarm');
COMMIT;
END;
END F_16;
thank you , in advance.
Your code has several issues I have addressed in the comments below. Note that transaction management is not discussed as it's not clear based on the question when commit/rollback should take place.
-- #1 use of explicit parameter mode
create or replace procedure f_16(p_trk_id in number, p_pilot_action in number) is
begin
-- #2 use of in
if p_pilot_action in (0, 1)
then
-- #3 unnecessary subquery removed
update controltow
set intention = p_pilot_action
where id = p_trk_id;
-- #4 use pl/sql implicit cursor attribute to check the number of affected rows
if sql%rowcount = 0
then
dbms_output.put_line('false alarm');
end if;
end if;
end;
Since you seem to be assigning pilot_action to Intention, I would do following:
create or replace
PROCEDURE F_16 (TRK_ID NUMBER, pilot_action NUMBER) IS
BEGIN
BEGIN
IF pilot_action IN (0, 1) THEN
-- if the only condition in subselect is the ID then use it directly
UPDATE "ControlTow"
SET "Intention"= pilot_action
WHERE "Id"=TRK_ID;
-- if there are more conditions than just the ID then subselect may be the way to go
--(hard to say without more information)
-- WHERE "Id" IN (
-- SELECT "Id" FROM "ControlTow" WHERE "Id"=TRK_ID AND ... )
ELSE
Null; -- do whatever you need in this case. Raise exception?
END IF;
EXCEPTION
WHEN NO_DATA_FOUND THEN dbms_output.put_line('False Alarm');
COMMIT;
END;
END F_16;
EDIT: As #user272735 said, there was room for more improvement on the code. Specifically rewriting the if condition to use in and simplifying the where clause (supposing Id is really the only condition to select rows to be updated).
I have a test table using a Microsoft SQL Server that is defined like this:
CREATE TABLE [dbo].[Table] (
[FirstName] NVARCHAR (255) NULL,
[LastName] NVARCHAR (255) NULL
);
There's just one row in the table with the values "person" and "man", respectively.
I'm trying to add a function that will update the values of that row but I keep running into this "[Microsoft][ODBC SQL Server Driver]String data, right truncation State code: 22001" error and I cannot figure out what the problem is. I've looked around and people say that it is caused by the data being too long to fit in the column but that's impossible because the string I'm trying to update with is only two characters, and as you can see in the table definition there is plenty of space for it.
I'm using a prepared statement for optimization purposes and the code creating it looks something like this:
const tString query("UPDATE \"" + tableName + "\" SET " + setClause + " WHERE " + whereClause + ";");
SQLHSTMT statement;
SQLAllocHandle(SQL_HANDLE_STMT, fSQLConnection, &statement);
SQLPrepareW(statement, (SQLWCHAR *) query.mWideStr(), SQL_NTS);`
The query string looks like this:
UPDATE "Table" SET "FirstName" = ?, "LastName" = ? WHERE "FirstName" = ? AND "LastName" = ?;
And then I am binding the parameters like this:
// We have our own string class that we use, which is where the mWideStr() and mGetStrSize()
// come from. mWideStr() returns a pointer to a UCS-2 buffer and mGetStrSize() returns the
// size in bytes.
SQLLEN pcbValue(SQL_NTS);
SQLUSMALLINT paramIndex(1);
// Call this for each parameter in the correct order they need to be bound, incrementing the
// index each time.
SQLBindParameter(statement, paramIndex++, SQL_PARAM_INPUT, SQL_C_WCHAR, SQL_VARCHAR, 255, 0, (SQLPOINTER) paramValue.mWideStr(), paramValue.mGetStrSize(), &pcbValue);
The first and second bound parameters are the new values which are both just "55", then third would be "person" and fourth would be "man".
Then to execute the statements it's just a call to SQLExecute():
SQLExecute(statement);
The call to SQLExecute() fails and then the error is generated and there is some more code that outputs the error message. As far as I can tell this should all be working perfectly fine. I have another database using Oracle that uses the exact same setup and code and it works without any issues, it's just SQL Server that's barfing for some reason. Is there something obviously wrong here that I'm missing? Does SQL Server have some weird rules that I need to add somewhere?
The SQLLEN pcbValue(SQL_NTS); variable being passed to SQLBindParameter() was going out of scope between binding the parameters and executing the statement, which means that some garbage data was being pointed to in the parameter binding. I also realized that you don't need to specify the last parameter. You can just pass NULL and it will act as if it is a nul-terminated string.
So the fix was to remove the SQLLEN pcbValue(SQL_NTS); variable and to just pass NULL to SQLBindParameter() for the last parameter instead.
Stupid mistake, but worth noting I suppose.
I have a table, that contains date field (let it be date s_date) and description field (varchar2(n) desc). What I need is to write a script (or a single query, if possible), that will parse the desc field and if it contains a valid oracle date, then it will cut this date and update the s_date, if it is null.
But there are one more condition - there are must be exactly one occurence of a date in the desc. If there are 0 or >1 - nothing should be updated.
By the time I came up with this pretty ugly solution using regular expressions:
----------------------------------------------
create or replace function to_date_single( p_date_str in varchar2 )
return date
is
l_date date;
pRegEx varchar(150);
pResStr varchar(150);
begin
pRegEx := '((0[1-9]|[12][0-9]|3[01])[.](0[1-9]|1[012])[.](19|20)\d\d)((.|\n|\t|\s)*((0[1-9]|[12][0-9]|3[01])[.](0[1-9]|1[012])[.](19|20)\d\d))?';
pResStr := regexp_substr(p_date_str, pRegEx);
if not (length(pResStr) = 10)
then return null;
end if;
l_date := to_date(pResStr, 'dd.mm.yyyy');
return l_date;
exception
when others then return null;
end to_date_single;
----------------------------------------------
update myTable t
set t.s_date = to_date_single(t.desc)
where t.s_date is null;
----------------------------------------------
But it's working extremely slow (more than a second for each record and i need to update about 30000 records). Is it possible to optimize the function somehow? Maybe it is the way to do the thing without regexp? Any other ideas?
Any advice is appreciated :)
EDIT:
OK, maybe it'll be useful for someone. The following regular expression performs check for valid date (DD.MM.YYYY) taking into account the number of days in a month, including the check for leap year:
(((0[1-9]|[12]\d|3[01])\.(0[13578]|1[02])\.((19|[2-9]\d)\d{2}))|((0[1-9]|[12]\d|30)\.(0[13456789]|1[012])\.((19|[2-9]\d)\d{2}))|((0[1-9]|1\d|2[0-8])\.02\.((19|[2-9]\d)\d{2}))|(29\.02\.((1[6-9]|[2-9]\d)(0[48]|[2468][048]|[13579][26])|((16|[2468][048]|[3579][26])00))))
I used it with the query, suggested by #David (see accepted answer), but I've tried select instead of update (so it's 1 regexp less per row, because we don't do regexp_substr) just for "benchmarking" purpose.
Numbers probably won't tell much here, cause it all depends on hardware, software and specific DB design, but it took about 2 minutes to select 36K records for me. Update will be slower, but I think It'll still be a reasonable time.
I would refactor it along the lines of a single update query.
Use two regexp_instr() calls in the where clause to find rows for which a first occurrence of the match occurs and a second occurrence does not, and regexp_substr() to pull the matching characters for the update.
update my_table
set my_date = to_date(regexp_subtr(desc,...),...)
where regexp_instr(desc,pattern,1,1) > 0 and
regexp_instr(desc,pattern,1,2) = 0
You might get even better performance with:
update my_table
set my_date = to_date(regexp_subtr(desc,...),...)
where case regexp_instr(desc,pattern,1,1)
when 0 then 'N'
else case regexp_instr(desc,pattern,1,2)
when 0 then 'Y'
else 'N'
end
end = 'Y'
... as it only evaluates the second regexp if the first is non-zero. The first query might also do that but the optimiser might choose to evaluate the second predicate first because it is an equality condition, under the assumption that it's more selective.
Or reordering the Case expression might be better -- it's a trade-off that's difficult to judge and probably very dependent on the data.
I think there's no way to improve this task. Actually, in order to achieve what you want it should get even slower.
Your regular expression matches text like 31.02.2013, 31.04.2013 outside the range of the month. If you put year in the game,
it gets even worse. 29.02.2012 is valid, but 29.02.2013 is not.
That's why you have to test if the result is a valid date.
Since there isn't a full regular expression for that, you would have to do it by PLSQL really.
In your to_date_single function you return null when a invalid date is found.
But that doesn't mean there won't be other valid dates forward on the text.
So you have to keep trying until you either find two valid dates or hit the end of the text:
create or replace function fn_to_date(p_date_str in varchar2) return date is
l_date date;
pRegEx varchar(150);
pResStr varchar(150);
vn_findings number;
vn_loop number;
begin
vn_findings := 0;
vn_loop := 1;
pRegEx := '((0[1-9]|[12][0-9]|3[01])[.](0[1-9]|1[012])[.](19|20)\d\d)';
loop
pResStr := regexp_substr(p_date_str, pRegEx, 1, vn_loop);
if pResStr is null then exit; end if;
begin
l_date := to_date(pResStr, 'dd.mm.yyyy');
vn_findings := vn_findings + 1;
-- your crazy requirement :)
if vn_findings = 2 then
return null;
end if;
exception when others then
null;
end;
-- you have to keep trying :)
vn_loop := vn_loop + 1;
end loop;
return l_date;
end;
Some tests:
select fn_to_date('xxxx29.02.2012xxxxx') c1 --ok
, fn_to_date('xxxx29.02.2012xxx29.02.2013xxx') c2 --ok, 2nd is invalid
, fn_to_date('xxxx29.02.2012xxx29.02.2016xxx') c2 --null, both are valid
from dual
As you are going to have to do try and error anyway one idea would be to use a simpler regular expression.
Something like \d\d[.]\d\d[.]\d\d\d\d would suffice. That would depend on your data, of course.
Using #David's idea you could filter the ammount of rows to apply your to_date_single function (because it's slow),
but regular expressions alone won't do what you want:
update my_table
set my_date = fn_to_date( )
where regexp_instr(desc,patern,1,1) > 0