Dear friends...

I am tired of try solve my problem. The System Tester not runs. The error it return is:

"A fatal error occured initializing System Tester"

Description:
Code = 80004003
Code meaning = Ponteiro inválido
Source = (null)
Description = (null)
Build: 9.0
File: ST_Database.cpp ...... [ ps: this file don´t exists in my computer!!!]
Line: 118
Error Type: 0
Error Code: 2

Any help ?

tks, Giovani

Hi GiOvanni, have you found a solution as yet? My system just crashed while running a system test, and I was getting the same error you have.
I did what DOC recommended (delete ST_Data.mdb file and reinstall MetaStock).
This fixed the problem. (Thanks DOC!)
It appears that the file "ST_RecordsetIterator.cpp" is inside the file ST_Data.mdb, but will not show up in a search.

Did you empty your recycle bin before re-installing MetaStock?
Maybe the old ST_Data.mdb was found and restored during installation?
Hope this helps.

Did you try to repair the database using Access prior to deleting and reinstalling? What about just deleting the records in the database or just deleting the database alone? Seems like reinstalling MS is pretty drastic and doesn't really answer the question about WHY the error occured in the first place. I suppose that there is a lesson here for all of us... make sure that you make a backup of key files immediately following a new installation. It is much easier to restore a file than reinstall everything.

Sorry about your troubles, but glad you found an answer... and shared it with the rest of us!

---

Traders' Consortium

Thanks g stockman, I have found how to repair the database from within XP using
Start>Control Panel>Administrative Tools>Data Sources.
Will use this if my system crashes again.
I'm not exactly sure what I was doing with the system tester at the time, to make it crash.
Cheers
charmeta

I'm getting this error also when running more complex and larger system tests.
(4 optimiziations with hundreds of equities => resulting in hundred thousands of tests)

Description: Invalid argument.

Build: 9.1
File: ST_RecordsetIterator.cpp
Line: 259
Error Type: 0
Error Code: 2Description: Invalid argument.

The funny thing is, usually the system test runs for a couple of hundred or thousand tests, and then it crashes.


I installed my MS 9.1 EoD on 3 different machines to check wether this would be something that is specific to one machine but now I've got this error on all the machines, and compacting the ST data file does not help on any of these.  Only very simple and small tests will run smoothly.

I believe there must be some fundamental error in the system tester and I hope Equis will come up with a solution patch quickly.

hey ben..... speakin of complex,  prof  d sornette is supposed to have moved over there by yall at eth....  he always had some complex crash views....

was wondering about your meta crashes.... have you checked the size of the 'st data' dat file..... also, is meta responding unusually slowly, such as when you have edited a indicator....   when you try to start the system tester, roughly how long does it take to come up after you have clicked on the icon.....h

thanks for the hint, I met prof sornette a couple of months ago at a conference, a great guy - also, his book is worthwhile reading.

- compacted the st data file several times, it usually was a little above 2'000'000 KB before compacting
- system tester is usually starting up veeery slowly when the problem is there.
- the simulations I run are currently very simple, optimizing 2 out of a combination of 4 Moving Averages, but I add on about 300 hundred stocks.


Also, do you have any experience with mining the st_data file through MS Access ?  because I would like to analyze the bulk results it in a more complex way than is currently possible with the System Tester.

Ben_Zurich:

Also, do you have any experience with mining the st_data file through MS Access ?  because I would like to analyze the bulk results it in a more complex way than is currently possible with the System Tester.

The ST_DATA.mdb is just a Jet database.  You can create a new front end in Access or any other application with ODBC and connect to the database for further analysis and testing.

You should NOT update or edit any records in the ST_DATA.mdb file from an external application unless you are very sure of what you are doing (too many people, even Forum members have come to grief trying to edit the database tables and data).

You can create a front-end in the ST_DATA.mdb itself, but this is not recommended.

Remember too, the results of each individual trade are NOT stored in the .mdb database, but in a directory called Results and are stored in .dta files which makes them about as useful as .... left-handed screwdrivers.


Hope this helps.

wabbit

---

"The question of whether a computer can think is no more interesting than the question of whether a submarine can swim."
Edsgar W. Dijkstra

 

MS: 6.52 EOD, 7.x EOD, 8.0 PRO, 9.2 PRO w/QC, 10 PRO w/QC C, 11 PRO w/QC & MDK
For custom MetaStock programming : http://www.wabbit.com.au
My SkyPE status :
My SkyPE account : wabbit.com.au

Hi,

I just deleted the ST_DATA file (the files were > 2 Gig on both workstations were I had installed it), and reinstalled Metastock over the old installation.

I now backuped the newly created ST_DATA file for future use, so maybe I could replace an erroneus file later with this one without having to reinstall MS completely.

Now my question is:

1) is this known that as soon as the ST_DATA file moves beyond a certain critical size (how much is it) it will start to crash?

2) since I know that I will test a variety of very large samples with optimizations right now, I imagine one (though cumbersome) temporary approach to that problem would be to backup the virgin ST_DATA file, run 1 or 2 highly complex simulations,  and then save that file under another name, in a different directory, and then restart with a virgin copy of the backup file?  So in the future, if I want to get back to these, I would just swap filenames?

3) is there a possilibity (as suggested in an another thread) to repair also a bigger file with Access' repair function in order to avoid this cumbersome procedure ?

4) Has Equis recognized this problem, and will something be done about it ???


Any input from more experienced users is highly welcome!

Thanks in advance.

2GB is the maximum file size allowed for Microsoft Access Databases.  This is an Access limitation, not a MetaStock limitation.  This is known, and it is recommended to back up the original MDB as a workaround (As you mentioned in point 2).

I just had a a bigger optimiziation run over night, with a total of 72576 steps.

It hang itself at step 46470 (64%) and I checked the new ST_DATA file which is again 2 Gig.


1)  Is there any approximate rule of thumb how many steps would be a maximum to stay below the 2 Gig limit?

2) and/ or is it known what does the size created depends on?
    - number of securities used
    - number of optimizations
    - date range / number of bars used per security?

I am getting errors both during installation as well as when trying to launch the system tester.

During install, FormOrg.exe fails with an abnormal program termination msg.

When I try to launch the system tester, I get this:

Description:
Code = 80004005
Code meaning = Unspecified error
Source = Microsoft JET Database Engine
Description = Could not find file 'C:\Program Files\Equis\MetaStock\ST_Data.mdb'.
Build: 10.0
File: ST_Database.cpp
Line: 118
Error Type: 0
Error Code: 2

I tried deleting the ST_Data.mdb file and reinstalling.  No joy.

The error above is complaining that the file should be in Metastock\st_data.mdb.  The actual file is in Metastock\temp\st_data.mdb. (I poked around and looked and there it was).

I don't see what I should do to change the location, as the only reference in the registry referrs to the program directory (one level above).

EDIT:

 

I just copied the files to the directory above and now it seems to work.  FormOrg.exe seems to work, too.