Curious about the accessible_columns and syscolumns tables.

I have a routine that exports the data in these tables to

a dos/windows text file.

When I look at the text file there are some foreign looking items.

can anyone explain the layout of these files?

My goal is to export 1 or these files to a text file.

Copy this file to a site. Run a routine to compare these

items with either the actual dict item or against their

syscolumns or accessible_columns table to check for

missing, altered dict items.

Do I need to 'cull' out the syscolumns before exporting?

Thanks in advance.

SYSCOLUMNS is a logical file. It doesn't really exist. It's just a reference of all the dictionary items in any file. The format is the same as the actual dictionary record.

The Sprezzatura Group

World Leaders in all things RevSoft

Well exactly. Thanks I guess.

What my question was in case I wasn't clear; was, could

I use this table (syscolumns) (assuming I had the maintain repos. checked for the app.) if

1) export to a text file

2) copy this to the remote site that has the app. I want to check

3) using a OI proc. read this file, each line would be id^item… and compare with the table dicts. attached.

That is the idea. What came about was, when doing (1)some of my items from syscolumns or accessible_columns (they seem the same) are 'not like the others (i.e. structure, lots of control characters, etc.)

Would this post merit a more detailed response?

Bruce,

Perhaps you should supply details first. Could you elaborate on that 'not like the others' thing? Examples?

- Oystein -

Sorry I was being as detailed as I knew at the time.

How about …

'What is in attribute of the accessible_columns table?'

And if I was to compare an item from accessible_columns, using the

item id, with the actual dict item would be a factor?

Thanks,

BC

The items in SYSCOLUMNS is the item in the dictionary. So, if you were to

dRec1=XLate( "VOC", "SIZE", "", "X")
dRec2=XLate( "SYSCOLUMNS", "VOC*SIZE", "", "X")

they would always be the same, since they are the exact same record.

Now, if you were to export the entire data into a text file and do the compare you suggested then that would validate one systems dictionary records against anothers.

However, you should not find differences in the SYSCOLUMNS representation and the actual dictionary representation. It's impossible, to the best of our knowledge.

The Sprezzatura Group

World Leaders in all things RevSoft

Sprezzatura,

But what's that field 51? SYSPROCS*DICT_EQUATES says it's for object code. What object code?

- Oystein -

I really do miss the fact that OI lacks the Date/Time stamp feature in Dict records for F and S type items. Wrote a global symbolic that looks like this:

K=COUNT(@RECORD,@FM)+1

XX=TRIM(@RECORD)

DTE=XX10,99

@ANS=ICONV(DTE,'D')

This returns the last date that a VOC item was compiled. So:

Select SOMEDICT with VOC_DATE ]= 'thisdate'

and voila .. extract the changes for update. Have a utility that allows for a saved list files to check for dict changes and so forth.

Don

All dictionaries have object code references in there.

For symbolics, it holds the object for the formula.

For data fields, it hold the object for @ANS=@Record.

The Sprezzatura Group

World Leaders in all things RevSoft

Yes, but in this case the remote version of the app. was

updated with an appbackup with no data so I think this is

the way to go!?

I did end up creating a text file but only copied attributes 1 thru 20.

Thanks for the reply.

View this thread on the forum...