Issues with record versioning

[Ricardo Oliveira]

Hi,

I'm having a couple of issues with record versioning:

1) I had to directly "force" versioning for tables (db.mytable._enable_record_versioning() for each table), although I have:

     auth.define_tables(username=True, signature=True)

     auth.enable_record_versioning(db, archive_names='%(tablename)s_archive', current_record='current_record')

     db._common_fields.append(auth.signature)

2) the user & time are not recorded on the modified_on, modified_by columns

Not sure where to go from here.

I would appreciate any help.

Best regards,

Ricardo.

[Ricardo Oliveira]

Hi,

I was able to fix problem 1).

Turns out that auth.enable_record_versioning needs to be after all the define_tables (found the answer on another post, but couldn't find it in the docs).

Still having issue 2). Has anyone seen this before? The solutions I saw on other posts didn't work for me.

Thanks.

[Anthony]

2) the user & time are not recorded on the modified_on, modified_by columns

How are you creating/updating the records?

[Ricardo Oliveira]

Hi Anthony,

Thanks for your reply.

I'm using the DAL's insert method for the creation. Eg:

db.table_x.insert(name='something')

For updates, I get the rows from the table first, change the values on the row object and then use the update_record() method for updating. Eg:

record = db(db.table_x.id == row_id).select().first()

record['name'] = 'something else'

record.update_record()

Thanks.

Ricardo.

[Áureo Dias Neto]

Hello,

You can do this, for update

db(db.person.id==1).update(name='Tim')

--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
You received this message because you are subscribed to the Google Groups "web2py-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ricardo Oliveira]

Thanks Áureo, good to know!

I use the other format because I do a few checks before doing the update, to see if the data is valid, but this short notation will no doubt be helpful in the future :)

Ricardo.

To unsubscribe from this group and stop receiving emails from it, send an email to web2py+un...@googlegroups.com.

[黄祥]

have the same situation but work fine, modified_on and modified_by is updated on the table. 

not sure why update_record on your side is not work.

best regards,

stifan

[Ricardo Oliveira]

I activated query logging but the only thing I can see is that when a change is made, the insert in the archive table has both the values (modified_on, modified_by) set to NULL.

Anyone else have any other ideas?

How can I further troubleshoot this issue?

TIA,

Ricardo.

[Anthony]

record = db(db.table_x.id == row_id).select().first()

record['name'] = 'something else'

record.update_record()

That's the one way you cannot do it. The problem is when you call .update_record() with no arguments, it takes all the existing fields in the record and uses them in the database update. It will therefore use the existing values of modified_by and modified_on. Typically, modified_by and modified_on get updated because they are excluded from the update call, which prompts the DAL to fill their values in automatically. But this mechanism breaks down when calling .update_record() with no arguments. We should probably add a note about this in the book.

Anthony

[Ricardo Oliveira]

Ah.. that should explain it then. In that case I'll have to do an update call with each of the update fields named individually.

That would be something nice to have in the book, I agree :)

Thanks Anthony, I'm going to test it now.

[Ricardo Oliveira]

Hi,

Just so it's clear for anyone having the same issue, Anthony's tip was right on target. All is working now.

Conclusion: you must use the update_record(field=new_value) format if you're using authentication, and not the update_record() format.

Thanks again to all that chipped in and Anthony for the solution.

Ricardo.