INFILE: transferring the source file from the source SAS server to the local machine

[khill]

Hello,

I'm trying to understand if I can avoid this behaviour. When using the INFILE function on a UNIX file, SAS first transfers the file to my local PC and then back to the server.

The source file is imported to the SAS server via PERL script to avoid my PC and it's relatively narrow network bandwidth.

The code looks like this;

DATA WORK.CDS_INFILE_CLEAN;
    LENGTH
        field1     $ 13
        field2     $ 24
     ;
    FORMAT
        field1     $CHAR13.
        field2     $CHAR24.
     ;
    INFORMAT
        field1     $CHAR13.
        field2     $CHAR24.
     ;     
    INFILE '/sas/saswork1/SAS_workA24500000D3B_g4u1113c/#LN00018'
        LRECL=337
        ENCODING="UTF8"
        TERMSTR=CRLF
        DLM='7F'x
        MISSOVER
        DSD ;
    INPUT
        field1     $CHAR13.
        field2     $CHAR24.
     ;
RUN;

Thanks,

John

[Doc_Duke]

John,

"When using the INFILE function on a UNIX file, SAS first transfers the file to my local PC and then back to the server."

That doesn't make any sense to me.  We have unix boxes and that doesn't happen on them.  You need to tell us more about the working environment or call tech support directly.

[That behavior is documented to happen with SAS datasets and EGuide, but there is also a documented approach to avoid it.]

Doc Muhlbaier

Duke

[khill]

Doc@Duke wrote:
[That behavior is documented to happen with SAS datasets and EGuide, but there is also a documented approach to avoid it.]

I can see the confusion. The dataset is imported via scripting, and then processed using the EG.  When I execute the sas program via UNIX script the import happens directly. However the program is executed within the EG I get the behaviour documented above.

Can you point me to the documented approach to avoid it...or is the approach to not use EG?

Thanks,

John