Information Analyzer

leomauer · Post by **leomauer** » Thu Feb 28, 2008 5:03 pm

Can Information Analyzer associate the Table definition in DataStage project with the sequential file on the node?
The IBM person told me is that I need to create ini file with metadata, but I do have it in DataStage already.
Also IBM person told me that for every sequential file I need to profile, I have to make an entry to .odbc.ini file and QETXT.INI file.
We have hundreds of files to analyze and it seems not practical. Especially considering IBM claim of unified metadata repository.
I appreciate any information/suggestion.

vmcburney · Post by **vmcburney** » Thu Feb 28, 2008 5:35 pm

Information Analyzer can only profile data through the new ODBC Connector stage - that's why you need an ODBC connection for each file. Databases are easier, you need just one connection for the entire database. There is an ODBC driver installed on my machine called "IBM Text File *.*" - can this driver be used to access any text file?

lstsaur · Post by **lstsaur** » Thu Feb 28, 2008 7:01 pm

leomauer,
You need only 1 entry in .odbc.ini and 1 QETXT.INI file for all of your sequential files. Of course, you must populate all the names and field descriptions of the sequential files. It's a pain. After that, all you can do on the sequential files is Column Analysis.

leomauer · Post by **leomauer** » Thu Feb 28, 2008 7:06 pm

We did connect to the file the way this IBM person described.
I would like to continue this discussion.
1. I still would like to know if the Table definition layout can be associated with flat file. To reiterate: I have a file and I have a table definition for this file entered into Analyzer project in DataStage. All I need is to point from the file in Analyzer to the Table definition.
2. The other thing that I would like to discover: is there any way to make this ODBC connections using Analyzer GUI.
All that takes to connect is to make the textual changes to .odbc.ini.
Wouldn't it be to insane to assume that somebody at IBM thought about it, and there is place in the GUI that allow to enter the required setting that are transferred to .odbc.ini file, and not the other way around?

leomauer · Post by **leomauer** » Thu Feb 28, 2008 7:11 pm

lstsaur wrote:leomauer,
You need only 1 entry in .odbc.ini and 1 QETXT.INI file for all of your sequential files. Of course, you must populate all the names and field descriptions of the sequential files. It's a pain. After that, all you can do on the sequential files is Column Analysis.

Thanks lstsaur.
Can you give me a sample of this single entry in both files that allow to describe all the files. It seems like I am not able to imagine how it looks like.

vmcburney · Post by **vmcburney** » Thu Feb 28, 2008 10:23 pm

The problem is that when you import a table definition into DataStage first you cannot use it in Info Analyzer - unless you use the Connector import from the Designer Import menu. The correct procedure is to import the table into Info Analyzer via the connector, profile it and then use that as a table definition in DataStage. This will give you a populated profiling summary tab when you look at the table properties from within the DataStage Designer.

Now I think this is inflexible - you should be able to profile any table definition within the Information Server as long as you have a connector to bring the data in - but that's the way it's currently designed. You should submit it to support as a defect/enhancement request.

Eventually I'm sure they will roll out more Connector stages and remove the plugin imports but at the moment we have two sets of metadata imports that do the same thing.

lstsaur · Post by **lstsaur** » Fri Feb 29, 2008 1:38 am

leomauer,
After I re-read your notes, I was wondering why you want to utilize the Table Definition of the DataStage to profile the seqential files?
To use Information Analyzer to profile your data, you create a new project, a Data Store, import meta data, and then associate the meta data with your project. Then, you will be able to do the Column Analysis, Table Analysis, Primary key, etc.
Anyway, the following is an example QETXT.INI file with multiple sequential files definitions should be coded:

[Defined Tables]
Categories.txt=Categories
Cust_Info.txt=Cust_Info
Division.txt=Division
[Categories]
FILE=Categories.txt
FLN=1
TT=Comma
Charset=ANSI
DS=
FIELD1=CategoryID,VARCHAR,11,0,11,0,
FIELD2=CategoryName,VARCHAR,14,0,14,0,
FIELD3=Description,VARCHAR,50,0,50,0,
FIELD4=OldCategoryID,VARCHAR,13,0,13,0,
[Cust_Info]
FILE=Cust_Info.txt
FLN=1
TT=Comma
Charset=ANSI
DS=
FIELD1=Cust_ID,VARCHAR,7,0,7,0,
FIELD2=SSN,VARCHAR,11,0,11,0,
FIELD3=Address,VARCHAR,28,0,28,0,
FIELD4=City,VARCHAR,12,0,12,0,
[Division]
FILE=Division.txt
FLN=1
TT=Comma
Charset=ANSI
DS=
FIELD1=DivisionID,NUMERIC,10,0,10,0,
FIELD2=DivisionName,VARCHAR,30,0,30,0,
........

leomauer · Post by **leomauer** » Fri Feb 29, 2008 12:35 pm

vmcburney wrote:The problem is that when you import a table definition into DataStage first you cannot use it in Info Analyzer - unless you use the Connector import from the Designer Import menu. The correct procedure is to import the table into Info Analyzer via the connector, profile it and then use that as a table definition in DataStage. This will give you a populated profiling summary tab when you look at the table properties from within the DataStage Designer.

Thanks vmcburney.
But I do not exactly know what you mean by "The correct procedure is to import the table into Info Analyzer via the connector".
The only way I see in GUI is to "Identify Next Level" which apparently goes to QETXT.INI file to grab the columns. Is that what you mean?
But that means that I need to put the column names and definitions into QETXT.INI file beforehand. I know it works. But I would like to know how to use the already existing table definition (imported from dsx file to DataStage Designer) if it is possible.

leomauer · Post by **leomauer** » Fri Feb 29, 2008 12:45 pm

lstsaur wrote:leomauer,
After I re-read your notes, I was wondering why you want to utilize the Table Definition of the DataStage to profile the seqential files?

Thanks for the sample. That exactly means what IBM person told us: there need to be as many entries into QETXT.INI file as there are files to profile.
To answer your question: our situation is - file comes from the source and the file layout comes from the Power Designer via MetaBroker right to DataStage Designer in the form of Table definition. So I thought that because we have the layout within the Information Analyzer suite, we have everything required to profile without typing the column definitions into INI file.