Extracting Metadata from a .DWG

[nicholas.amphlett]

Hi there,

I have been given a task by work to figure out a way of extracting metadata from a .dwg so that the information can be inputted into an application that is being developed.

It is my understanding that a .DWG is purely an image with attributes and there is no metadata, however there is text information in each .dwg that i need.

I have 15,000 drawings to sift through and manually viewing each one and noting the information down i.e drawing number, location etc etc would be an extremely tedius job that would take far too long.

Does anyone have any ideas how i can do this? preferably with a batch application.

I have had a look on google and have found a few applications that would be able to extract information (attributes) but none that extract the information that i actually need.

Any help would be greatly appreciated,

Thanks very much

Nick

[rkmcswain]

What information are you trying to catalog? You mentioned the drawing number, location, etc.

Would something like this work?
http://www.dotsoft.com/demos/tpdwgcatgen.htm

[nicholas.amphlett]

What information are you trying to catalog? You mentioned the drawing number, location, etc.

Would something like this work?
http://www.dotsoft.com/demos/tpdwgcatgen.htm

The application could potentially work, just trying to get a free trial to give it a go,

Thanks very much for your help,

Nick

[jaberwok]

It is my understanding that a .DWG is purely an image with attributes and there is no metadata, however there is text information in each .dwg that i need.

Just FYI, a DWG file is actually a database, not an image.

[nicholas.amphlett]

I have tried to get in contact with that company about a trial of there product but i have had no response,

In response to the question about what information i require, when you open up a dwg drawing there is borders around the drawing itself and on the right hand side (i think) there is information such as drawing number, location, client address etc etc - that is the information that i need to get from 15,000 dwg's.

Does anyonbody else have any other ideas please?

Thanks again,

Nick

[jaberwok]

When acad opens a dwg file it reads the contents of the (database) file and generates graphical elements (lines, circles, text) accordingly.

You could check to see if the information has been saved as metadata by right-clicking, one at a time, on each of a few files, selecting Properties and seeing if there is any information on the Summary and Custom tabs. I doubt that there will be.

You could save each file to the DXF format, open them in a text editor (Wordpad) and search for the info you want but that would take forever even for a few files.
You need to automate that process (in effect if not it fact) and the best place to start would probably be in the programming forums.

[nicholas.amphlett]

When acad opens a dwg file it reads the contents of the (database) file and generates graphical elements (lines, circles, text) accordingly.

You could check to see if the information has been saved as metadata by right-clicking, one at a time, on each of a few files, selecting Properties and seeing if there is any information on the Summary and Custom tabs. I doubt that there will be.

You could save each file to the DXF format, open them in a text editor (Wordpad) and search for the info you want but that would take forever even for a few files.
You need to automate that process (in effect if not it fact) and the best place to start would probably be in the programming forums.

Thanks alot for your response, i have tried your recommendation and it is definetly not an efficient way of doing this.

Keep the ideas coming though guys! Really appreciate it!

Thanks again

Nick

[dotson]

I have tried to get in contact with that company about a trial of there product but i have had no response

Just noting that at the time you posted this it had been roughly 21 hours since your email.

In response to the question about what information i require, when you open up a dwg drawing there is borders around the drawing itself and on the right hand side (i think) there is information such as drawing number, location, client address etc etc - that is the information that i need to get from 15,000 dwg's.

Assuming ALL these drawings are consistant with that information in a block insert it would fairly easily be harvested, possibly by tools built into AutoCAD. If it's random text it would provide more of a challenge. That is the first determination you will need to make.

[nicholas.amphlett]

Hey Dotson,

Further to our telephone conversation i have checked with the department that deals with DWG's. They have advised that the majority of the drawings are consistent however the text that is required is "completely separate from the rest of the drawing" - they are not attributes - with that in mind they also advised that there would be no efficient way within AUTOCAD to get this information and the only solution they say would work is to hire some temps to go through the drawings and manually input the text needed.

Thanks again,

Nick

[jaberwok]

Hey Dotson,

Further to our telephone conversation i have checked with the department that deals with DWG's. They have advised that the majority of the drawings are consistent however the text that is required is "completely separate from the rest of the drawing" - they are not attributes - with that in mind they also advised that there would be no efficient way within AUTOCAD to get this information and the only solution they say would work is to hire some temps to go through the drawings and manually input the text needed.

Thanks again,

Nick

If Terry Dotson says it can be done, it can be done.