Subject | Re: PDOXPUMP |
---|---|
Author | hwelmore2002 <hwelmore@kimbanet.com> |
Post date | 2003-01-30T18:52:28Z |
--- In IB-Conversions@yahoogroups.com, Helen Borrie <helebor@t...>
wrote:
script is added for the necessary field translation when data is
pumped. I'm still getting the invalid variant error with the latest
download.
generated as they were in the 2002 verion of PDOXPump.
single digit deg C and snow!
have helped isolate the problem. I know the ultimate script must be
what I need for the pump.
Guide but could not find it. Will be glad to download and study if I
could locate it.
Helen, I realize you have clients to take care of and I appreciate
the time you have taken to figure a solution to my problem. My
migration of data to Firebird is under no time contraints yet so
whenever a fix or response is approprite for you is fine with me.
Thanks again
Wayne
wrote:
> At 03:07 PM 29/01/2003 +0000, you wrote:happening
> >Helen, I've run into a different problem now. Here is whats
> >now.IB_Domain_Map
> >
> >1. Using IB_SQL I've created a database (dialect 3)
> >2. Executed IB_Domain.SQL
> >3. Browsed the database and the Country, IB_Domain, and
> >tables are created.Dialect
> >4. Executed PDOXPUMP
> >5. Logged into database
> >6. On Metadata and Data Pump screen, checked Do Datapump and
> >3 boxes.the
> >7. BDE login and loaded table to pump
> >8. Added to scrip the single primary index (field of VARCHAR(3))
> >9. Refreshed Domains
> >10. Selected each field with domain types and selected the proper
> >domain type in the combobox.
> >11. Clicked add to script for table.
> >
> >At this point I get the error message 'Invalid Variant Type
> >Conversion'.
> >In the table script window is the following:
> >
> >Create Table Vehicle(
> > Veh Varchar(3) Not Null
> >
> >Out of curiosity, I clicked the Add table script again and received
> >the same error msg above. And now in the table script is:
> >
> >Create Tabel Vehicle(
> > Veh Varchar(3) Not Null
> > P_Date D_Date Not Null
> >
> >Clicking the add table script once more added to the table script
> >create table vehicle( line and all fields from the pdox table.in
> >
> >However there was nothing in the Domains and Generators window.
> >Naturally I can't pump the table without the Domains being created
> >the script and I forget the exact systax to use.for any
> >Anyway, can you send us a fix?
>
> There is no "fix" for this. The tool will create the statements
> generators that you indicate are needed and it does offer somesuggested
> default domain definitions for some types. The window is there toprovide
> feedback on generators and domains that are created at the time yourequest
> them and, at the right time, it will add those newly-createdstatements to
> the script. If the default domains don't fit your data, you needto edit
> them manually.This is where I intended to ask for a fix so the domains creation
script is added for the necessary field translation when data is
pumped. I'm still getting the invalid variant error with the latest
download.
> Once they have been added to the script, any furtherdoesn't
> editing of them needs to be done in the main script. The program
> go back and compare the statements in the main script with what'sleft
> behind in the contributing windows. It also doesn't later parsethrough a
> saved script and extract your CREATE DOMAIN and CREATE GENERATORstatements
> back into the feedback window - otherwise, potentially, they wouldbe added
> again to the main script. If you are working on a saved script,look in
> the window in the bottom right corner (I think) - where the latestversion
> of the script (whether previously saved or not) should appear.The CREATE DOMAIN and CREATE GENERATOR statements are not being
>
generated as they were in the 2002 verion of PDOXPump.
> I suspect the invalid variant type errors may be coming from mycode
> calling the AsDateTime method on a value that resolves to NULL .The
> "As... methods work by casting variants to types. NULL isn't avariant,
> and I know I handled this in the original code using NULLVAR, whichworked
> properly in IBO3.IB 6.0
>
> The tool was built with IBO 3 in Delphi 5 and tested with IB 5.6,
> and the first Firebird beta. When I recompiled it the other day, Iused
> the latest IBO4 and Delphi 6. A lot has happened to IBO, Delphiand
> Firebird since then, so it would be worth reviewing the variantbecause
> handling. Perhaps IBO4 has changed the way NULLVAR is interpreted,
> Delphi 6 had some pretty scary bugs concerning variants and Jasongenerally
> does try to make workarounds for Delphi bug...I can't promise toget to it
> straight away, though, as I have urgent customer work to do rightSorry it is so hot there. Wish we had some of it here what with the
> now. And, dammit, it's 38 deg C in the shade here today! :-|
>
single digit deg C and snow!
> I'm a bit dubious about you apparently piggybacking re-runs of theapp upon
> one another. The ultimate script - the one that you decide to usefor the
> pump - has to be a complete script, not bits and pieces...it's notusing
> ALTER <object> on subsequent runs to override anything you havealready
> created in the main script or in the database.fine, but
>
> You have to get all the metadata right before you start pumping
> data. Trial runs to test stuff and find likely problem areas are
> you need to be prepared to delete stuff and re-run it until you areAs mentioned before, I did this our of curiosity and thought it might
> satisfied that your metadata will work with the incoming data.
>
have helped isolate the problem. I know the ultimate script must be
what I need for the pump.
> As the notes say, this utility really does require that youunderstand what
> will happen when the metadata script runs and be capable of editingit to
> fit with what you know about the actual input and output data. Themain
> benefit of the tool (provided you understand what's going on) isthat it
> makes it possible for you to create the metadata and pump the datawithin
> one UI. It doesn't pretend to be bomb-proof - it's a merger ofseveral
> little utility progs I'd developed for myself over the years forpdox to ib
> conversions, using the BDE on both sides of the boat - hence theneed to
> keep the steps in strictly the right order.by its
>
> When I started using IBO for the data-pumping, I was so astonished
> speed that I suggested to Jason that it might be useful to work myutility
> routines into some sort of integrated migration tool, to help outin cases
> like yours. Pdox2IB was the result. It would have been nice tohave time
> to make it "newbie-proof" but there wasn't, and it's not.beta
>
> If you need more info about the Dialect 3 types, read the IB 6.0
> Migration Guide, which you can download from the IBPhoenix site.Went to IBPhoenix.com and searched for the IB 6.0 beta Migration
Guide but could not find it. Will be glad to download and study if I
could locate it.
> It soundsDefinition
> as if you could do with the Language Reference and the Data
> Guide as well. :-)And point me to those documents and I'll be glad to study them also.
>
> heLen
Helen, I realize you have clients to take care of and I appreciate
the time you have taken to figure a solution to my problem. My
migration of data to Firebird is under no time contraints yet so
whenever a fix or response is approprite for you is fine with me.
Thanks again
Wayne