URL classification function: Parameters and examples - Simple Log Service

The URL classification function groups similar URL request paths, assigns them a common pattern, and generates a corresponding regular expression for each group to simplify URL categorization. You can then use the query results for ETL operations.

Note

The URL classification function is only available in the China (Beijing) and China (Shanghai) regions.

Syntax

select url_classify(url_path varchar);
select url_classify(url_path varchar, weight long);

Input parameters

Parameter	Description
url_path	The URL request path.
weight	The weight for classification.

Output parameters

Parameter	Description
url_path	The original URL request path.
api_path	The generalized API endpoint pattern.
regex_tpl	A regular expression template that matches the pattern.

Output

     url_path                          |     api_path                             |         regex_tpl
-------------------------------------+------------------------------+-------------------------------------
 /gl/balance/666398186799140           | /gl/balance/*                             | \/gl\/balance\/[0-9].+
 /gl/glaccount/30579281472076          | /gl/glaccount/*                         | \/gl\/glaccount\/[0-9].+
 /gl/balance/709016207098025           | /gl/balance/*                            | \/gl\/balance\/[0-9].+

Example
- Query statement
```
* | select url_classify(uri, num) from (select uri, COUNT(*) as num from log group by uri limit 1000)
```
- The query returns three columns: url_path (the original request path), api_path (the classified API path pattern, where an asterisk (*) replaces variable parts), and regex_tpl (the corresponding regular expression template). For example, the original path /v1/task/20200403_064500_63933_w69w5.2.28/results/17/1 is classified as /v1/task/*/results/17/1, and its corresponding regular expression template is \/v1\/task\/.*\/results\/17\/1.