topic Re: Proc import - ERROR: Failed to connect to the Server in SAS Enterprise Guide

Proc import - ERROR: Failed to connect to the Server

Autotelic — Mon, 02 Oct 2017 09:47:09 GMT

SO: Windows 10

SAS EG: 7.1 (7.100.0.1966) (64-bit) Edit: SAS Base 9.2

Proc import out= WORK.auto1 datafile= "path\file_name.xlsx" 
            DBMS=excelcs
            replace;
            sheet="auto"; 
Run;

When trying to execute the above code, I get the error:

ERROR: Failed to connect to the Server: .

I'm not connected on any server:

The file I'm trying to import was downloaded from this example and my code is based on the code in the link.

What's going on? How can I use proc import to import this excel?

Re: Proc import - ERROR: Failed to connect to the Server

LinusH — Fri, 08 Sep 2017 11:42:14 GMT

Excelcs requires a PC Files Server i think.
Try other engine, or check the status of your pc files server.

Re: Proc import - ERROR: Failed to connect to the Server

Autotelic — Fri, 08 Sep 2017 11:49:35 GMT

Alas I have no idea of what you're talking about. What is another engine? How can I check the status of my pc files server?

Re: Proc import - ERROR: Failed to connect to the Server

Kurt_Bremser — Fri, 08 Sep 2017 12:31:59 GMT

The excelxs engine is meant for access to old (pre-xlsx) files from non-Windows platforms. It needs a "SAS PC Files Server" running on a Windows machine that uses the Windows-only Microsoft-supplied conversion module.

Since you have a local install, you can try to use dbms=xlsx instead (no extra server needed, but SAS 9.4 required)

Re: Proc import - ERROR: Failed to connect to the Server

Autotelic — Fri, 08 Sep 2017 13:15:51 GMT

I'm on SAS 9.2 (I just added this information to the post).
I assume this is why I do a manual import, it converts to csv first.

Re: Proc import - ERROR: Failed to connect to the Server

ballardw — Fri, 08 Sep 2017 17:49:42 GMT

@Autotelic wrote:
I'm on SAS 9.2 (I just added this information to the post).
I assume this is why I do a manual import, it converts to csv first.

I routinely convert Excel files to CSV prior to reading them with a data step because proc import has to guess about variable types and lengths. In most cases repeated files of supposedly the same the layout and similar content will get changes in lengths of character variables or types of variables when using import causing headaches when combining the data. I head that off by the data step reading a known structure and setting the variable types and lenghths based on the agreed description of the data I should be getting.