I am trying to set up a third party MySQL as a Linked Server on SQL Server 2008.
When I try to access the data via a simple OpenQuery below I get an error.
SELECT * FROM OpenQuery(SERVERNAME, 'SELECT * FROM table')
The error message:
OLE DB provider 'MSDASQL' for linked server 'SERVERNAME' returned data that does not match expected data length for column '[MSDASQL].tablename'. The (maximum) expected data length is X, while the returned data length is Y.
NOTE:(where X > Y)
I've tried with another MySQL table that I built myself and it works fine. Therefore I assume the problem might be with the MySQL source.
EDIT: After digging a little deeper into the MySQL data I found several "invalid" dates such as zero dates. These could be a reason, however I have no way to change the data source so I must find a way to convince SQL to ignore it.
I recall something form previous versions of SQL where you could switch to "Lazy Schema Validation" to force SQL not to check this. However this option appears to be gone in 2008.
Any ideas how I could make this work?
Related
Brief background of the system:
Using an SSIS package to move data from SQL Server to MySQL.
I use an OLE DB connection to connect to SQL Server.
I use an ODBC Unicode connection to connect to MySQL.
Most of the system works, but I am having a date conversion issue when passing an SSIS Variable into an Execute SQL Task. It attempts insert a date into a table on the MySQL database. I get the following error when trying to enter a date of {1/1/2010 12:00:00 AM}.
"[Execute SQL Task] Error: Executing the query "Insert into ETLDate values(?)" failed with the following error: "[MySQL][ODBC 5.3(w) Driver][mysqld-5.7.22-log]Incorrect datetime value: '-10165-58296-9719 16613:13824:2136' for column 'LastETLDate' at row 1". Possible failure reasons: Problems with the query, "ResultSet" property not set correctly, parameters not set correctly, or connection not established correctly."
The date is consistently '-10165-58296-9719 16613:13824:2136' so I know the value is getting passed correctly, it's just converted or encoded differently. Here is my setup for the parameter.
Some final notes:
I've tried all the other date data types for the parameter and I get the same issue.
I've tried with just a direct insert statement using the parameter and not a stored procedure and I get the same issue.
All data flow tasks that move data (which includes dates) from the SQL Server DB to the MySQL database work fine.
If create a similar stored procedure and table on the SQL Server DB (with updated parameter settings for the OLE DB connection) it works fine.
Any one know why the date is getting passed incorrectly or have a work around for this issue?
I have created a SSIS package in order to transfer datas from ACCESS to SQL SERVER.
Source > SQL Command from "mdb" file joinning two tables
Destination > Flat table in SQL Server
I'm performing the JOIN in the source SQL Command because of the number of records in the ACCESS tables (~500k).
I tried to use SSIS join but it take ages doing the ORDERING before JOIN.
While running package in VS2010, it works great.
But after deploying and executing the package on my SQL Server 2014 the following error occurs.
No column information was returned by the SQL command.
Returned validation status "VS_NEEDSNEWMETADATA"."
I'm pretty sure my SQL command is correct (Working in VS and the preview button in the editor show me records).
I tried to disable ValidateMetadata but the same error still occurs, but at execution this time.
In the SQL Server 2014 I have other packages calling ACCESS data (But without join) and it works properly.
Thanks for your help,
Q.
ValidateMetadata is (generally) a good thing.
This error is being caused by the metadata on your source or destination (it's unclear from your question) being different.
At a guess, at least one of the columns in your SQL2014 database is of a different datatype (or length, or is nullable etc.) - there is a difference either way.
This little gem has kept me going for a wee while now and I'm still not much further forward.
Environment
SSMS 2012 SP1 (target) (Server 2012 64 bit)
MySQL (source)
Linked server using ODBC 5.3 64bit Unicode driver
I am trying to port the tables from MySQL onto SSMS.
A majority of the tables work fine....until I hit a column that is cast > 8000 chars.
As I was having an issue I tried to just select the offending Columnm
Running a simple
SELECT RogueCol FROM LinkServer...MyTbl
returns the error
OLE DB provider "MSDASQL" for linked server "LinkServer" returned message "Requested conversion is not supported.".
Msg 7341, Level 16, State 2, Line 1
Cannot get the current row value of column "[MSDASQL].RogueCol" from OLE DB provider "MSDASQL" for linked server "LinkServer".
So I then tried
SELECT * FROM OPENQUERY(LinkServer, 'SELECT RogueCol FROM MyTbl')
This returns the same error
I'm totally confused as I cannot seem to even view the data in SSMS, let alone SELECT it into a table (I have a script that will cast the fields according to some rules based on their original MySQL castings)
As I say, for a majority of other fields its not an issue.
The MySQL casting of the RogueCol is varchar(32767)
The settings enabled on the MSDASQL Provider is:
Nested Queries
Level Zero Only
Allow Inprocess
Supports 'Like'
So far I've tried a raft of things which I've 'discovered' on various forums including:
CASTING, CONVERTING, LEFT, LTRIM(RTRIM()) on the actual call.
Tried also restricting it to a TOP 1 record (LEN = 45, Doesn't have any white space or foreign characters in the string)
Set up the ANSI version of the ODBC driver
Various settings within the ODBC driver (Allow Big Result sets, changing Character sets, Limit Columns to 32 bit, Pad char with space among others)
Can I call upon your collective knowledge and request some further suggestions please as this driving me nuts?
Many thanks
UPDATE
So guys and girls, no responses
I ended up having to talk to the DBA responsible for the MySQL DB and asked him to alter the casting of the RogueCol from varchar(32767) to TEXT
Job jobbed
I need to copy data from a remote mySQL database into MSSQL Server 2008 database using SSIS 2008 package. I have some Timestamp fields in the mySQL database giving me problems. When l excluded the Timestamp fields l managed to copy the data nicely. On the OLE DB Destination Input - Input Columns the field which is Timestamp on mySQL database is showing as having data type DT_Bytes.
I am getting the error An OLE DB record is available.
Source: "Microsoft SQL Server Native Client 10.0" Hresult: 0x80040E21 Description: "Multiple-step OLE DB operation generated errors. Check each OLE DB status value, if available. No work was done."
[OLE DB Destination [784]] Error: Cannot create an OLE DB accessor. Verify that the column metadata is valid.
My select statement to the mySQL source has got a select case to handle 0 dates it is like;
SELECT case modified_date when day(modified_date) = 0 then '1990-01-01 00:00:00' else modified _date end as modified from mySQLTableName
The problem is definitely with your data types. Open up your Destination component for editing. Go the mapping tab. Hover over each column mapping, first the source column, then the destination. The tooltip will show you what the data types of each column are. One of them doesn't match.
It's most likely your timestamps field, but it might not be, so I won't speculate.
In any case, to make them match, add a Data Transformation component in between your source and destination, and recast the offending source column to the data type you want for inserting into SQL Server. (If the recasting also causes an error, it'll be a lot more helpful one than this one!)
I have an SSIS package set up to export data from a SQL Server 2008 R2 table to a MySQL version of that table. The package executes however, I am getting about 1% of the rows failing to be exported.
My source connection uses the SQL statement
SELECT * FROM Table1
all of the columns are integers. An example of a row which is exported successfully is
2169,2680, 3532,NULL, 2169
compared to a row which fails
2168,2679,3532,NULL, 2168
virtually nothing different that I can ascertain.
Notably, if I change the source query to only attempt the transfer of a single failing row - ie.
SELECT * FROM Table1 WHERE ID = 2168
then the record is exported fine - it is only when part of a select which returns multiple rows that it fails. The same rows fail the export each time. I have redirected error rows to a text file which displays a -1071610801 error for the failing rows. This would apparently translate to:-
DTS_E_ADODESTERRORUPDATEROW: "An error has occurred while sending this row to destination data source."
which doesn't really add a great deal to my understanding of the issue!
I am wondering if there is a locking issue or something preventing given rows from being fetched or inserted correctly but if anyone has any ideas or suggestions on what might be causing this or even better how to go about resolving it they would be greatly appreciated. I am currently at a total loss...
Try to setup longer timeout (1 day) ot the mysql (ADO.NET) destination.
Well after much head scratching and attempting every work around that I could come up with I have finally found a solution for this.
In the end I switched out the MySQL connector for a different driver produced by devArt -dotConnect for MySql and, with a few minor exceptions (which I think I can resolve) all of my data is now exporting without error.
The driver is a paid for product unfortunately but in the end I'd have taken out a new mortgage to see all those tasks go green!