This article is obviously a follow up to Permission Problems post I published previously. The other day I was using the User Permissions Report (RPUSER), I noticed something amiss. I had recently deleted a couple of educational and testing companies which many people used. However, like many other areas of M2M, the records are not properly deleted and cleaned up.

This becomes a problem when you run the User Permissions Report because it lists every single company, every user associated with it, and the exhaustive list of screens with their permissions. It includes every user and company even after deleting them from within M2M. This slows down the report and clutters the results.

Below is the code one could use in Visual FoxPro to clean this up. Please read my Standard Disclaimer before trying any code from this site. I’m going to stress this several times because of an issue that came up as a result of the orphaned records post. Stay tuned for details.

Set excl on – You must have exclusive access to perform this.
OPEN DATABASE X:\m2mdata\util\util.dbc – Replace X with the network drive letter assigned to M2M.
Use utacc
Copy to utaccOLD
Brow for fccompid = ‘XX’ – where XX is the number of the M2M Company for which you intend to remove users and rights. Make sure you do NOT do so on your live M2M Company. This will show the records that will be deleted.
Delete for fccompid = ‘XX’ – Remember, don’t use this on your live data.
Pack – This will make the deletion permanent.
Close all
Quit.

Please remember to test this on a test server, and if you feel uncomfortable contact M2M Support or a qualified consultant such as myself for help.

Any questions?

As I was preparing to post another script, this time with an update statement, I started to insert my standard disclaimer. However, as more and more of my posts will include SQL code, I thought I would just dedicate one post for the disclaimer and then link to it. So, without further adieu…

Important: All information on this site is provided “as is” for informational purposes only.

In particular, all code samples are provided “as is” without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

If you should decide to use any of the code taken from this site, make sure you test it thoroughly on a test company, or better yet a test server, before you ever use it on your live data.

That isn’t to say that I don’t want to help you, I help people all the time. Just be careful when you apply any code to your server or database. I do my best to research everything I post, but ultimately you are responsible for your own database.

Keep that in mind when you run code from my site or any site you find on the internet because….

One of my crappy college jobs was in market research. Yes, I was one of those annoying guys in the mall holding a clipboard. I would approach you, politely ask you about your soda preferences (for the enlightened folks we call it Pop), and then take you back into the “store” to do a taste test on several of them. Kind of like the Pepsi Challenge.

It required a few minutes, but it was general painless and we’d send you away with samples of free products. Well, I don’t want to know whether you prefer Coke or Pepsi, but I have what I think are some amazing ideas for reporting on M2M. I need your assistance to determine whether people can/will actually use them. So, in the same vein, please assist me by answering some questions.

How large is your largest M2M database? To determine the database size, simply open Enterprise Manager, navigate to your database, right click on it, and choose Properties. You should see the size of the database on the first page shown (General).
Note: There is a poll embedded within this post, please visit the site to participate in this post's poll.

Is disk space an issue for you? If I told you I could simply your reporting life by automating most of the reporting process, but it would involve duplicating at least parts of your database, would this cause you an issue? Keep in mind that this process could double or triple your disk usage.
Note: There is a poll embedded within this post, please visit the site to participate in this post's poll.

Please paste the following script into Query Analyzer or SQL Server Management Studio. It will report the size of many of the primary tables in your M2M Database. Please respond by copying and pasting the grid into the comments area, like people did with the Orphaned Records Game.

DECLARE @TableName VARCHAR(100)    --For storing values in the cursor
 
--Cursor to get the name of all user tables from the sysobjects listing
DECLARE tableCursor CURSOR
FOR 
SELECT [name]
FROM dbo.sysobjects 
WHERE  OBJECTPROPERTY(id, N'IsUserTable') = 1
and [name] in 
('slcdpm',
         'somast',
         'soitem',
         'sorels',
         'qtmast',
         'qtitem',
         'inmast',
         'inonhd',
         'invend',
         'armast',
         'aritem',
         'apmast',
         'apitem',
         'jomast',
         'joitem',
         'pomast',
         'poitem',
         'lamast',
         'ladetail',
         'shmast',
         'shitem',
         'glmast',
         'glitem',
         'rcmast',
         'rcitem')
FOR READ ONLY
 
--A procedure level temp table to store the results
CREATE TABLE #TempTable
(
    tableName VARCHAR(100),
    numberofRows VARCHAR(100),
    reservedSize VARCHAR(50),
    dataSize VARCHAR(50),
    indexSize VARCHAR(50),
    unusedSize VARCHAR(50)
)
 
--Open the cursor
OPEN tableCursor
 
--Get the first table name from the cursor
FETCH NEXT FROM tableCursor INTO @TableName
 
--Loop until the cursor was not able to fetch
WHILE (@@FETCH_STATUS >= 0)
BEGIN
    --Dump the results of the sp_spaceused query to the temp table
    INSERT  #TempTable
        EXEC SP_SPACEUSED @TableName
 
    --Get the next table name
    FETCH NEXT FROM tableCursor INTO @TableName
END
 
--Get rid of the cursor
CLOSE tableCursor
DEALLOCATE tableCursor
 
--Select all records so we can use the reults
SELECT tablename AS [TABLE],datasize AS [SIZE]
FROM #TempTable
ORDER BY tablename
 
--Final cleanup!
DROP TABLE #TempTable

Thank you for participating, and stay tuned for updates on my new reporting ideas.

My girlfriend really loves the Amazing Race Show, which of course means that I have to “love” it too. It’s a reality TV show where teams compete in contests all over the world with a one million dollar prize at the end. Anyway, in one episode of last season the contestants were in Dubai. The challenge was to weigh out the exactly $500,000 worth of gold and the value was constantly fluctuating. They were given the value of gold per ounce. The trick was to realize that 16 ounces are in a pound and for the sake of this example the price of gold is $1089.87 per ounce. If you divide $500,000 by the price per ounce you can figure out how many ounces you need. Then divide that number by 16 to determine how many pounds and ounces for the final weight. Simple right? Well, remember that the price of gold was changing every minute or so. Think you can do it in under a minute?

Go ahead and grab a piece of paper and try. I’ll wait.
::Humming the Jeopardy Theme::

What? You can’t do it? Everyone learns how to do it, it’s just simply long division.

You’ve forgotten how to do long division with decimals? I realized the same thing as my girlfriend and I were laughing at the contestants. I have forgotten how to do long division. This is ironic because I used to teach math. I am a SQL Programmer, which obviously involves math. I’m embarrassed to admit that I actually Googled how to do long division with decimals.

I can almost hear you saying, “Well stupid, use a calculator or the computer you are working on.” Well, some of the teams had calculators and got the answer. The teams that didn’t couldn’t keep up with the constantly changing price of gold and had to do an alternate challenge.

What is the point?

Why do we use calculators and computers? It’s simple, they are fast. What is the likelihood that a million dollars will be on the line for a long division problem? What is the likelihood that I will be without some kind of gadget that calculates?

In this case, it is a no-brainer. I may never have to do long division again. However, other technologies are not so clear-cut. Since moving to Dallas nearly two years ago I have used a Garmin Nuvi GPS and I love those things. I type in the address I’m looking for and it gives me turn by turn directions. It helps me find stores, restaurants, and I only use a small subset of the features of this device.

Here’s the problem. Since I use this thing everywhere I go, I have virtually no understanding of the geography here. Remember when you were a little kid and you were driven everywhere? Do you remember what it was like to start driving yourself? I actually had to “find” places that I had been to tons of times because I never had to know where they were before. In essence, this is my situation now. If someone asks me for directions to my place, I’m fairly clueless.

Can you still write cursive (besides your signature)? I really have to work at it and I hate filling out forms. Have you noticed that your spelling has degenerated since the advent of spell check? Mine has.

What the heck does this have to do with SQL?

How many of us could effectively program without intellisense? I learned Basic waaaay back in the day and really had to learn it. When discussing development tools I often hear veteran programmers laugh that you can code in notepad, but who would? We did. We learned more thoroughly too. When mistakes are painful, we make less of them. I truly think that people learning now, don’t learn as thoroughly because of intellisense.

Furthermore, as I’ve mentioned before, when I write T-SQL I do so using a program called TOAD. I really like the program and it greatly improves my productivity. However, if I open Query Analyzer or SQL Server Management Studio, I still have to Google many of the functions to get the syntax right. Is using TOAD, or another product doing myself a disservice in the long run?

What do you think?

A Made2Manage customer called me with a serious performance problem with one of his custom reports. After I looked at it, I realized that the inventory changes in M2M Version 6.0 were the culprit. To give you some idea of his problem, the query in question ran took a few seconds before the migration and then jumped to a whopping 2.5 minutes afterward.

Since M2M Support won’t support custom reports, who you gonna call? Ghostbust….. Well no, actually me.

What was the problem with M2M 6.0?

I remembered from Consona Connect 2008 that the inventory master table(INMAST) was essentially replaced with a sub table INMASTX and converted into a view. However, at that time, nobody could/would tell me why. Well, take a look at the create statement for the view:

/****** Object: View [dbo].[inmast] ******/
USE [M2MDATA01];
GO
SET  ANSI_NULLS ON;
GO
SET  QUOTED_IDENTIFIER ON;
GO
 
-- Create new view and instead of trigger based on the renamed table.
CREATE VIEW dbo.inmast AS
  SELECT *
       , dbo.GetItemOnHandQuantity(fac, fpartno, frev) AS fonhand
       , dbo.GetItemInspectionQuantity(fac, fpartno, frev) AS fqtyinspec
       , dbo.GetItemNonNetQuantity(fac, fpartno, frev) AS fnonnetqty
       , dbo.GetItemInProcessQuantity(fac, fpartno, frev) AS fproqty
       , dbo.GetItemOnOrderQuantity(fac, fpartno, frev) AS fonorder
       , dbo.GetItemCommittedQuantity(fac, fpartno, frev) AS fbook
       , dbo.GetItemLastIssueDate(fac, fpartno, frev) AS flastiss
       , dbo.GetItemLastReceiptDate(fac, fpartno, frev) AS flastrcpt
       , dbo.GetItemMTDIssues(fac
                            , fpartno
                            , frev
                            , GETDATE())
           AS fmtdiss
       , dbo.GetItemYTDIssues(fac
                            , fpartno
                            , frev
                            , GETDATE())
           AS fytdiss
       , dbo.GetItemMTDReceipts(fac
                              , fpartno
                              , frev
                              , GETDATE())
           AS fmtdrcpt
       , dbo.GetItemYTDReceipts(fac
                              , fpartno
                              , frev
                              , GETDATE())
           AS fytdrcpt
  FROM dbo.inmastx
GO

I’ll briefly describe what is happening here. The view loads every field from the base table IMASTX and then completes 12 separate User Defined Functions. I described how to use UDF’s here.

Why did Consona make this change?

I’d say it’s a good bet that they did so because they’ve always had problems keeping inventory numbers synced with other tables, especially with the Inventory On Hand table (INONHD). That’s why there are so many utilities to recalculate these manually. It’s important to note that M2M violates the rules of database normalization by keeping duplicates of information all over its database, and this is just one more example of it.

Great, what is the problem?

The problem, as I eluded to before, is that this kills performance, or at least the querying performance. Each one of those UDFs makes a separate call to the database, converting one select statement into many for each row.

Consider the following profile traces. When I execute the following query against the underlying table:

SELECT *
FROM inmastx
WHERE fpartno = '123456'
AND   frev = '001'

Profiler returns 13 rows with a total run time of 29 Milliseconds or .029 seconds.
The next trace involves the same query against the INMAST view:

SELECT *
FROM inmast
WHERE fpartno = ‘123456’
AND   frev = '001'

Profiler returns 49 Rows with a total runtime of 198 Milliseconds or .198 seconds.

What’s the big hairy deal? The second query is 1/5th of a second anyway.

Ah, very true, but how many parts do you have in your item master? I ran the following queries against a database with approximately 7000 part numbers. Keep in mind that you should never “Select *” from anything, but I am doing it for simplicity.

SELECT * FROM inmastx requires 1 MINUTE 22 seconds. 
SELECT * FROM inmast requires 3 minutes 4 seconds.

This performance hit above is bad, but it actually gets worse. What happens if you are only retrieving fields that aren’t calculated? For example, what if you only want the part number, revision, and product class? This is common because some modules, such as shipping, do not include the product class so you must link back to the item master to retrieve it. Consider the following:

SELECT FPARTNO, FREV, FPRODCL FROM m2mdata01..inmast

Because INMAST is a view, all of those columns are calculated anyway and the query churns 6 times as long regardless.

So, what do I do about it?

Well, you’re going to have to edit your custom reports that deal with inventory in order to migrate, or consult with someone like me to do it for you. Essentially, you need to query the underlying table INMASTX for your data and then if you need On Hand Quantity, calculate it yourself with a sub-query. Here’s an example from above.

SELECT *
, dbo.GetItemOnHandQuantity(fac, fpartno, frev) [OnHand]
FROM inmastx 
WHERE fpartno = '123456'
AND   frev = '001'

This isn’t a perfect solution, but in this way you can call any of the UDF’s used in that view without calling all of them.

Well smarty pants, if this is so bad, what would you do?

First, it’s Mr. Smarty Pants to you. Seriously though, I have reservations about the upgrade path with M2M. To the best of my knowledge, Consona is not making substantive database changes in Version 7. In my opinion, they should take the opportunity to normalize the database (or at least significant portions of it) and enforce referential integrity while migrating to the .Net platform. I think they should drop VFP entirely, train users to use SQL Reporting Services for reports, and start fresh in version 7.

What do you think?