如何通过Python直接调用SAS存储过程服务器以高效获取数据？

Absolutely, you can directly invoke the SAS Stored Process Server without going through the SAS Stored Process Web Application layer—this is exactly the right approach to cut down on overhead and enable efficient data transfer to Python via the _WEBOUT file. Since you’ve already ruled out SASPy (correctly, as it relies on Workspace Servers for interactive sessions), here are the most practical methods to achieve this:

1. Direct REST API Call (Simplest & Most Common) #

The SAS Stored Process Server exposes a REST endpoint that you can hit directly from Python using libraries like requests. This skips the web application’s UI layer entirely and communicates straight with the server.

Step-by-Step Example: #

First, confirm your SAS server’s endpoint (typically http://<your-sas-server>:<port>/SASStoredProcess/do for SAS 9; for SAS Viya, it’s usually http://<viya-server>/SASJobExecution). Then use this Python code:

import requests
from requests.auth import HTTPBasicAuth
import pandas as pd
from io import StringIO

# Configure your SAS server details
SAS_SERVER = "your-sas-server-hostname"
SAS_PORT = 8080  # Default port for SAS 9; adjust for Viya
STORED_PROC_PATH = "/Your/Stored/Process/Location/YourSP.sas"
USERNAME = "your-sas-username"
PASSWORD = "your-sas-password"

# Build request URL and parameters
api_url = f"http://{SAS_SERVER}:{SAS_PORT}/SASStoredProcess/do"
request_params = {
    "_program": STORED_PROC_PATH,
    "_output_type": "text"  # Match your _WEBOUT format (text/csv/json)
}

# Send authenticated POST request
response = requests.post(
    api_url,
    params=request_params,
    auth=HTTPBasicAuth(USERNAME, PASSWORD)
)

# Process the _WEBOUT output
if response.status_code == 200:
    # Example: Parse CSV output into a pandas DataFrame
    df = pd.read_csv(StringIO(response.text))
    print("Successfully retrieved data:")
    print(df.head())
else:
    print(f"Failed to call stored process: {response.status_code} - {response.text}")

2. Optimize _WEBOUT for Faster Transfer #

To maximize efficiency, format your SAS stored process’s _WEBOUT output into compact, machine-readable formats:

• JSON: Use proc json to output structured data directly to _WEBOUT:

  /* SAS Stored Process code */
  filename _webout temp;
  proc json out=_webout nosastags pretty;
      export sashelp.class / keys;
  run;
  

• Compressed CSV: Reduce payload size with gzip compression:

  /* SAS Stored Process code */
  filename _webout temp gzip;
  proc export data=sashelp.class outfile=_webout dbms=csv replace;
  run;
  

  In Python, you’d decompress the response before parsing:

  import gzip
  from io import BytesIO
  
  decompressed_data = gzip.decompress(response.content)
  df = pd.read_csv(StringIO(decompressed_data.decode('utf-8')))
  

3. Key Configuration Checks #

Before you start, make sure:

• Your SAS administrator has enabled direct API access to the Stored Process Server
• The server’s firewall allows traffic from your Python environment to the SAS port
• You’re using the correct authentication method (Basic Auth, OAuth2, or SAS Token—check with your SAS admin)

内容的提问来源于stack exchange，提问作者submit