The source of this it turns out was actually a period in the Publication name, as we [...]]]> We had merge replication in SQL Server 2005 set up that was working fine, except when trying to view conflicts, the Conflict Viewer would give this error: “RowGUID is neither a DataColumn nor a DataRelation for table summary. (System.Data)”.

The source of this it turns out was actually a period in the Publication name, as we had a “6.1” as part of the name. Other special characters would probably also cause trouble.

The only way to fix this was to drop the replication, resolve known conflicts manually with a data compare, recreate the replication, and reinitialize. After that, we could use the Conflict Viewer to resolve conflicts successfully.

This was a frustrating issue since the replication was allowed to be created with the period and worked completely fine, only the Conflict Viewer experienced trouble and it gave an error that did not accurately describe the issue.

The moral of the story is: don’t create a publication with a period in the name.

This example was created based on work in this thread: http://www.sqlservercentral.com/Forums/Topic802508-1672-1.aspx

Let’s say you have a data set like so:

and for each distinct “id”, you want [...]]]> Sometimes we need to rollup string values into lists of comma separated values. Using “FOR XML PATH” per column that you want to rollup can provide a way to do this.

This example was created based on work in this thread:
http://www.sqlservercentral.com/Forums/Topic802508-1672-1.aspx

Let’s say you have a data set like so:

and for each distinct “id”, you want a list of its distinct fruits and cities.

If you put a FOR XML PATH statement for each column you want to rollup like so:

...
----------fruit---------
,STUFF((SELECT DISTINCT ', ' + NULLIF(fruit,'')
FROM tableToCSVRollup t2
WHERE t1.id = t2.id
FOR XML PATH(''),TYPE
).value('.','VARCHAR(MAX)')
,1,2,'') AS fruit
...

and then GROUP BY id, you’ll have a nice aggregated set. It will even put it in alphabetical order and remove NULL and empty string values for you.

Complete code example:

;WITH tableToCSVRollup (id,fruit,city)
AS (
SELECT 101,'Apple','Cleveland'
UNION ALL SELECT 101,'Apple','Pittsburgh'
UNION ALL SELECT 101,'Banana','Pittsburgh'
UNION ALL SELECT 102,'Grape','Cleveland'
UNION ALL SELECT 102,'Melon','Cleveland'
UNION ALL SELECT 103,'Melon','Pittsburgh'
UNION ALL SELECT 103,'Melon','Cleveland'
)

SELECT
id
----------fruit---------
,STUFF((SELECT DISTINCT ', ' + NULLIF(fruit,'')
FROM tableToCSVRollup t2
WHERE t1.id = t2.id
FOR XML PATH(''),TYPE
).value('.','VARCHAR(MAX)')
,1,2,'') AS fruit
----------city----------
,STUFF((SELECT DISTINCT ', ' + NULLIF(city,'')
FROM tableToCSVRollup t2
WHERE t1.id = t2.id
FOR XML PATH(''),TYPE
).value('.','VARCHAR(MAX)')
,1,2,'') AS city
------------------------
FROM tableToCSVRollup t1
GROUP BY id
ORDER BY id

Results:

——————————————

Table A is the Header, Table B a sub-header related to table A via an ID, and Table C related to Table B via an id. [...]]]> Today, as I was working on a project, I found an opportunity to use the For Loop Container. The basic setup, was three tables, A, B, C for simplicity.

Table A is the Header, Table B a sub-header related to table A via an ID, and Table C related to Table B via an id. I wanted to loop through the Id’s in table B one at a time due to the amount of data. I pulled in TableAId, MinTableBId and MaxTableBId into an object variable, Using For Each Loop Container, I wanted to loop through each TableAId and then loop through each TableBId, pulling records from Table C one at a time. I decided to drop a For Loop Container into the For Each Loop Container. Initially, there was no issue, everything looped through. As the iterations increased, my laptop stopped responding. I was able to see the memory was sitting at 500 megs…then 600… eventually topping out around 850 megs (There is only 1 gig of RAM on the laptop I use at work.). By this time, I had to kill the task.

I thought maybe it was a fluke, so I restarted (when in doubt restart right=)). Cleaned out the tables and ran again. Same result.

After ripping out the for loop, and using a different technique, 150 megs of RAM was consumed. I plan on doing some more tests to see if I can find more useful information for everyone, but I wanted to post this as a caution to be careful with a For Loop Container. This is with SQL 2005 SP2.