topic Re: How To Add Root XML Tags To Large File in SAS Programming

How To Add Root XML Tags To Large File

shl007 — Thu, 14 May 2020 14:36:42 GMT

Hi - I have some starter code below I'm trying to apply to add <root> and end </root> tags to a large XML file. Two questions:

1) is there a way to specify a DIFFERENT output file name?

2) Any ideas on how to get around a workspace “disconnected” error? Is there an alternate approach I can use to put <root> and </root> tags in the file?

  data _null_;
     infile 'c:\output\temp.xml' end=eof;
    input;
      file "c:\output\temp1.xml";
         if _n_=1 then do;
        put "<root>";
        end;
            put _infile_;
         if eof then do;
       put "</root>";
   end;   
run;

Re: How To Add Root XML Tags To Large File

shl007 — Thu, 14 May 2020 17:06:33 GMT

#1 - I answered my own question via the code; the output filename is already specified.
#2 - still TBD

Re: How To Add Root XML Tags To Large File

ballardw — Thu, 14 May 2020 20:57:08 GMT

Anything involving "disconnected" is likely not related directly to your code.

Since the code provided works on a small file there is nothing wrong there. However you may need to consider LRECL for your actual file if it has real long "lines" (> 32767 ). This may happen if your xml has freeform text fields that contain extensive text.

Another approach might be to create two very small text files one with the "<root>" and the other

"</root>". Then use a system tool to concatenate the 3 files.

I have to assume with your comment about "disconnected" that you are attempting to modify a file on a remote file system somewhere. It may be that you need to copy the file to a local location, run your code creating a local updated version and then upload the modified version back to the problematic location.

If that approach is not feasible then you have to diagnose conditions causing the disconnect which may be well out of the scope of many of us on this forum.

Re: How To Add Root XML Tags To Large File

shl007 — Thu, 14 May 2020 21:15:41 GMT

I'm thinking some of the issue is the unstructured nature of my XML file, as far as using a data step to insert <root> and </root>. Instead of a system tool, I'm hopeful I can just import the separate files into separate datasets and concatenate them thru data step to get the <root> and </root> inserted at beginning and end of the file, respectively. Let me know if you can think of any pitfalls with this approach other than its lack of elegance ...

The wrinkle here is, I've modified the code to use a PIPE command. So I may be importing multiple files at once. This may be tricky (i.e., the <root> and </root> would need to be inserted for EACH file).

Re: How To Add Root XML Tags To Large File

ballardw — Thu, 14 May 2020 22:51:51 GMT

Have you been to successfully just read in ONE of these files?

You say the issue is disconnecting. If you get disconnected from your data source I don't see how using an approach to read multiple of these files is going to make anything better?

Re: How To Add Root XML Tags To Large File c:\output\temp.xml

ChrisNZ — Thu, 14 May 2020 23:26:04 GMT

To go around the timeout issue, you could run an OS command. Save the two extra lines and run:

copy header.txt+c:\output\temp.xml+footer.txt c:\output\temp1.xml