Update table according to matches found in second table - sql

Question

Having a heavy brain freeze here, all help is welcome.
Background info: working on a search function that should return the number of occurrences of one or multiple (user provided) keywords based on matches in another table.
The first table will actually be a temporary table inside a table function but for the sake of simplicity, let's make it a regular table.
CREATE TABLE search_query {
keyword VARCHAR (30),
occurrences INT
}
That "table" will contain the (sanitized) keywords the user provides. For example:
keyword occurrences
HOW 0
TO 0
CODE 0
THIS 0
"occurences" is 0 by default.
My second table contains a record for each significant keyword a product can have.
CREATE TABLE product_keywords {
productId INT,
keyword VARCHAR(30)
}
Sample data:
productId keyword
12 HOW
12 NOT
12 CODE
13 RANDOM
13 THIS
13 CODE
What I'm trying to do is build a query that will UPDATE the "occurrences" field in search_query based on keyword matches with product_keywords.
In this example, after the query runs, search_query should contain
HOW 1
TO 0
CODE 2
THIS 1
My efforts so far stranded because for the life of me, I can't figure out how to build the JOIN necessary for the update. I'm about here:
UPDATE destination_table
SET destination_table.occurrences = JN.matches
FROM search_query AS destination_table
INNER JOIN [where I freeze] AS JN
WHERE ....
Any help appreciated! Thanks, Chris

Answer 1

You need a subquery to return the matches and join it to the table:
UPDATE s
SET s.occurrences = g.counter
FROM search_query s INNER JOIN (
SELECT s.keyword, COUNT(*) counter
FROM search_query s INNER JOIN product_keywords p
ON p.keyword = s.keyword
GROUP BY s.keyword
) g ON g.keyword = s.keyword
See the demo.
Or for readability you could use a CTE:
WITH cte AS (
SELECT s.keyword, COUNT(*) counter
FROM search_query s INNER JOIN product_keywords p
ON p.keyword = s.keyword
GROUP BY s.keyword
)
UPDATE s
SET s.occurrences = c.counter
FROM search_query s INNER JOIN cte c
ON c.keyword = s.keyword
See the demo.
Results:
> keyword | occurrences
> :------ | ----------:
> HOW | 1
> TO | 0
> CODE | 2
> THIS | 1

Answer 2

A correlated subquery with UPDATE might be the simplest and fastest approach:
update search_query
set occurrences = (select count(*)
from product_keywords pk
where pk.keyword = search_query.keyword
);
For performance, you want an index on product_keywords(keywords).