Changing SQL Server Collation After Installation

I had a necessity to change the collation of newly created SQL server instance from CI_AS to CP1_CI_AS.

To brief you about why this was needed, I wanted to test upgrade GP from version 2010 RTM to 2010 SP3. In order to that, I created a new SQL server instance with collation SQL_Latin1_General_CI_AS. I restored GP 2010 RTM backups on to this newly created server. But later realised that GP databases that I got for 2010 RTM was of collation SQL_Latin1_General_CP1_CI_AS.

There were two options. Either I must change the collation of databases to CI_AS or I must change the server collation itself to CP1_CI_AS. Later seemed to be a better option (for some reason, I wanted to retain the collation on the databases). I was left with changing the collation of server instance.

That’s when I got this link: Changing SQL Server Collation After Installation

I was worried about the whole process but this post saved me loads of time and sweat. Changing the collation was a breeze.

Anybody out there who is looking for the best possible method to change server collation, above is the post you would want to read.


CBM Cash Receipt AA Journal Issue & Resolution

Couple of days back, a user reported an issue. When she tried to open CBM Cash Receipts window, following message appeared:

Screen Shot 2014-11-25 at 5.48.52 PM

For all other users, when they posted any Cash Receipt, an Analytical Accounting Journal report would open with some entries on it; particularly with 5 cash receipts. User, who reported the issue, confirmed that the records shown on AA journal report were posted by her and only after posting them, the above error message appeared.

Since the report is AA journal, I first checked the AA Sub Ledger tables (AAG20000 series) and confirmed that the entries shown on the AA journal report are present without any issues. If I get rid of the above error message, things would be normal and this particular user can continue with her entries. Task on hand is to find out AA journal report tables and see the stuck/orphaned records and clear them.

Issue: Records got stuck on AAG50000 (AA Sub Ledger Report Header), AAG50001 (AA Sub Ledger Report Distribution) and AAG50002 (AA Sub Ledger Report Assignment) tables.

Resolution: Remove them. Simple. What you MUST note here is that all records (including genuine ones) would be present on these tables. You MUST note down AA Sub Ledger Header ID of the ones that are stuck and use below query to delete records.

DELETE AAG50000 WHERE aaSubLedgerHdrID IN (<aaSubLedgerHdrID>,<aaSubLedgerHdrID>,...) 
DELETE AAG50001 WHERE aaSubLedgerHdrID IN (<aaSubLedgerHdrID>,<aaSubLedgerHdrID>,...) 
DELETE AAG50002 WHERE aaSubLedgerHdrID IN (<aaSubLedgerHdrID>,<aaSubLedgerHdrID>,...)

Replace <aaSubLedgerHdrID> with respective AA Sub Ledger Header ID of the stuck records.

NOTE: Always take a backup of records before you delete.


Simulate DEX_ROW_ID in a SQL View Using ROW_NUMBER() – #MSDYNGP

I have a requirement in which I have to access a SQL view from within my customisation dictionary, in order to create a custom lookup for users to select a value based on an Extender form and an Extender lookup. Easiest option is to create an Extender view (which in turn creates a SQL view for us).

Now, this is the view that I am suppose to refer to from my custom dictionary. Dexterity allows us to refer to any SQL object by simply create a TABLE definition and mention the SQL object (table or view) name as the physical name.

Dex View

Everything looks perfect till you actually see below error messages at runtime:

Dex View - Error 1

Dex View - Error 2

Error message is quite obvious; you do not have DEX_ROW_ID in that SQL view that you are referring to. Every single Dexterity table must have DEX_ROW_ID at the backend. It cannot afford to not have one.

So how am I going to resolve this? By simply adding a record number dynamically to the SQL view created by Extender. How to do that? By adding the T-SQL function ROW_NUMBER(). This is how I achieved it:



Definition of ROW_NUMBER() can be found here: ROW_NUMBER (Transact-SQL).

A simple yet powerful SQL function has given me the power to do what I wanted in no time. Oh, and my custom lookup referring to this view is working like a charm. Users are happy and so am I.


syExcelReports Table & SQL Server Native Client 10.0 – Thanks Aaron Berquist

While upgrading our GP from GP2010 SP3 to GP2013 SP1, DYNAMICS database upgrade stopped unexpectedly with GP crashing without any information for me to debug.

But it stopped exactly at table syExcelReports and crashed. Without spending anymore time, I just went and checked my ODBC DSN setup, as explained by Aaron Berquist on his blog High Dynamics Range. Expectedly, the ODBC DSN driver was not pointing to SQL Server Native Client 10.0.

For some reason, GP fails to process upgrade further if your DSN driver is NOT native client 10.0 or above.

Those who upgrade your GP to latest version, please check DSN before you start database upgrade using GP Utilities.



There is an amazingly simpler yet highly informative post up on TECHNET portal, which proved very handy this morning.

My systems admin reported to me that our SharePoint server ran out of storage for some reason unknown to him, but he could identify that it was SQL Server program files folder which consumed 80% of storage.

I immediately knew what could be the reason. It’s those crazy log files problem. To truncate and clear these overeating log files, I had to use this command: DBCC SHRINKFILE.


Once I cleared out all unnecessary file space from these log files, I had to setup the File Growth and Maximum File Size (shown in the screenshot beside):

Never for a DB log file, we should setup the file growth as “In Percent”. This would prove disastrous over a period of time. Instead, always set it up to “In Megabytes” and enter a least sensible value; in my case I had set it up to 5MB.

For a DB Log file, 2GB should be reasonable amount of file space to hold the logs. To learn more about Log files, read this post on TECHNET: The Transaction Log (SQL Server).

We must understand one thing, that shrinking the DB log has got it’s own impact.


Implicit Conversions Causing Deadlocks in SQL Server

Quite an interesting but crucial post up there on Sans SQL blog.

The post explains that Implicit conversions in SQL Server could actually trigger a deadlock. And there are definitions for each important term (such as deadlock, implicit conversion, etc.) with an example.

I believe this post would be useful for those who work extensively on SQL Server.


T-SQL Function for Proper Case Format

We do not have a built in T-SQL function to convert any string (or a statement) to Proper Case format, also known as Title Case. This I feel is a very simple feature that SQL Server could have provided us, but missing even in it’s latest version, SQL Server 2012.

I thought I would highlight some links which would be useful for T-SQL programmers in our community:

  1. David Wiseman’s Solution
  2. Jeff Smith’s Solution
  3. Pinal Dave’s Solution

There are many more solutions, but above are noteworthy.