Pattern Matching Functions
Used with the MATCH Clause, the Vertica pattern matching functions return additional data about the patterns found/output. For example, you can use these functions to return values representing the name of the event or pattern that matched the input row, the sequential number of the match, or a partition-wide unique identifier for the instance of the pattern that matched.
Pattern matching is particularly useful for clickstream analysis where you might want to identify users' actions based on their Web browsing behavior (page clicks). A typical online clickstream funnel is:
Company home page -> product home page -> search -> results -> purchase online
Using the above clickstream funnel, you can search for a match on the user's sequence of web clicks and identify that the user:
- Landed on the company home page.
- Navigated to the product page.
- Ran a search.
- Clicked a link from the search results.
- Made a purchase.
For examples that use this clickstream model, see Event Series Pattern Matching.
GROUP BY and PARTITION BY expressions do not support window functions.