I just made a discovery, this user sent me his DB to test the update on. He has been one of my users for 10 years. I noticed that the ssnum and Sp_ssnum columns were exactly the same for 60% and that stopped in Feb 07.
Back in Feb 07 I sent an update that updated him from 6.4 to 7.5 using the same Unload method. That 07 update caused the same problem 3 years ago but I never noticed it. The only thing that has saved me is the field that got screw up is not used for anything important but what if it was an important column that got screw up? At this point I have zero faith in my Unload method of updating users, If the update / Unload / Load runs without errors how would you know this happened? It was just dump luck I decided to make an unload file in the final DB and compare them. This is not practical for out of state users. If I discover why I will pass it on Marc From: A.G. IJntema Sent: Wednesday, April 28, 2010 4:55 PM To: RBASE-L Mailing List Subject: [RBASE-L] - RE: Another Strange problem Marc, Probably it does not help. But I have had a similar experience with loading XML files. In this file there were column names which were more or less the same. In one way or another Rbase was confused about this and the data were not loaded correctly. I changed the names a little bit and then it was all right. Maybe you should give it a try To give you an idea, what I have done: - <REPLACE> <CONTAINING From="ID=" To="IDA=" /> </REPLACE> So my suggestion is to change the name sp_ssnum in something else, if possible and see what happens Tony From: [email protected] [mailto:[email protected]] On Behalf Of MDRD Sent: woensdag 28 april 2010 23:00 To: RBASE-L Mailing List Subject: [RBASE-L] - RE: Another Strange problem Jan and Jason I am using 7.6 30305, I know not the latest build, I had some screens display in a different place on the screen and got a note from RBase that it would be fixed on the next update. It appears that the output file is created correctly but when you load it back into a blank DB the ssnum overwrites whatever is in the sp_ssnum files. There are no defaults on either field, and the data looks OK in the output file, the problem seems to be when you run the file to Load the table. DELETE ptinfo.dat OUTPUT ptinfo.dat UNLOAD DATA + FOR ptinfo USING custnum,f_name,l_name,address,city,state,ptzip,hm_phone,+ wk_phone,b_date,bdaycard,ssnum,sex,typclaim,married,midint,empstat,reff_by,+ sp_phone,sp_ssnum,statemen,work_acc,mail_lis,other_acc,finc_cla,pt_part,+ ins_name,relation,I.................... thanks Marc From: jan johansen Sent: Wednesday, April 28, 2010 3:01 PM To: RBASE-L Mailing List Subject: [RBASE-L] - RE: Another Strange problem Marc, It kind of looks like your 12th column has a NULL FORMAT of '000000000' applied to it. Jan -----Original Message----- From: "MDRD" <[email protected]> To: [email protected] (RBASE-L Mailing List) Date: Wed, 28 Apr 2010 14:55:08 -0500 Subject: [RBASE-L] - RE: Another Strange problem This is getting the best of me. I cut part of the Output files to show the before data and after data. At this point I am afraid to update any of these users. Even if I take my clean DB Unload the data Using Column, delete all my data, then run the file to load the tables I get this new data in the table. So in this case the DB is an exact copy I am trying to load the data back into. Original 19000',10/04/1989,'Y','000000000','F','5','S',,'2',,,,'1','N','Y','N','IN',+ after Load 19000',10/04/1989,'Y','000000000','F','5','S',,'2',,,' 000000000','1','N','Y'+ Any other suggestions, Marc From: A.G. IJntema Sent: Wednesday, April 28, 2010 1:00 AM To: RBASE-L Mailing List Subject: [RBASE-L] - RE: Another Strange problem Maybe you can do it in another way. This is the way how I have done it succesfully several times First I define per table a variable with all the columns in it , which you like to convert -- CAO SET VAR V_CAO_COL = 'CAO_ID,JRTL_ID,CAO_volgnr,CAO_mnd,CAO_pro,CAO_vb,CAO_min,CAO_max' Open old database Then unload the data as ascii -- CAO OUTPUT CAO.PDN UNLOAD DATA FOR CAO USING &V_CAO_COL AS ASCII OUTPUT SCREEN Open new database To be sure the new table is empty I run the statement: DELETE ROWS FROM CAO Finally the data are loaded into the new table as ascii SET SEMI=NULL SET SEMI= '^' LOAD CAO FROM CAO. PDN USING & V_CAO_COL ^ NOCHECK ^ NONUM SET SEMI=NULL SET SEMI= ';' In this way you are in control of the conversion and you can add some checks and balances in the routine. For instance count the number of rows in the old and new table, you easily can avoid calculated fields and so on . With defining the columns as a variable you are sure the structure is the same in the old and new version. Another advantage is you can optimize the new database in the way you like to have it and in the end there is a clean situation, which is the same for everyone. The only thing you have to keep in mind is the order of loading the data into the new database. PK’s and FK’s Hope this helps. Tony IJntema From: rbas...@rbase .com [mailto:[email protected]] On Behalf Of MDRD Sent: woensdag 28 april 2010 0:00 To: RBASE-L Mailing List Subject: [RBASE-L] - Another Strange problem Hi I am trying to update my users to 7,6 and I thought I would Unload Data for Table Using ..... for all the data tables then load them into a blank 7.6 DB. Some of my users haven't updated my app for years and and I thought this would be a good way to make sure they had the exact same structure I had. I have dozens of out of state users. I did not get any errors but to double check the data I did a Unload All on the original DB and the new 7.6 DB after I loaded the data. Then I compared the 2 files to see if the data matched and it did Not? In the first example you can see the 2 was changed to a 3... very strange. In the second example, don't worry I changed the name and numbers with fake data But you can see on the second line a -0- was replaced with a SS number from the first line You can see that they have "dell" for the name so I wonder if that is the problem. Also I noticed sometimes they have\ Smith, Jr in the last name on some customers but sometimes there is no strange " or , in the data and the data after the load is different. This makes me real nervous about sending out my update. neb',-0-,'244999999','3','N','N','N','IN',0,'Same','1','318 Valley Stream+ --- original file neb',-0-,'3 44999999','3','N','N','N','IN',0,'Same','1','318 Valley Stream+ -- after Load 9444',-0-,11/12/1967,'Y','243089000','F','5','M','L','1','Sam O''dell',+ --- original file -0-,-0-,'3','N','Y','N','IN',0,'Eddie, Kryan','2','3400 Fixcraft Road','Char+ 9444',-0-,11/12/1967,'Y','243089000','F','5','M','L','1','Sam O''dell',+ --- after Load -0-,'243089000 ','3','N','Y','N','IN',0,'Eddie, Kryan','2','3400 Fixcraft Roa+ Any suggestions? Marc
