Changed pending email query structure to improve speed

[howlinghuffy]

Q	A
Bug fix?	Y
New feature?	N
Automated tests included?	N
Related user documentation PR URL	N
Related developer documentation PR URL	N
Issues addressed (#s or URLs)	#2449 #3826 #3358
BC breaks?	N
Deprecations?	N

Description:

This fixes a performance bottleneck when loading the Channels -> Emails page with a large number of contacts in Mautic. The old code uses an inefficient NOT EXISTS query, and the new code uses a more efficient NOT IN query.

See an analysis of the different types of queries in MySQL here

To summarise, the query is going from the form:

SELECT  l.*
FROM    t_left l
WHERE   NOT EXISTS
        (
        SELECT  NULL
        FROM    t_right r
        WHERE   r.value = l.value
        )

to

SELECT  l.*
FROM    t_left l
WHERE   l.value NOT IN
        (
        SELECT  value
        FROM    t_right r
        )

It should be noted that the NOT EXISTS and NOT IN queries are slightly different in how they handle NULL values. However, this is irrelevant in our scenario, as the lead_list_leads.lead_id field to which we are comparing the subqueries IDs are set to NOT NULL, so the results will always be equivalent.

Tests with 500,000 contacts show a 200x speed improvement for this query, from 200.5s down to 1.04s.

Steps to reproduce the bug:

1. Import 500,000+ customer emails
2. Add at least one email and assign it to a segment with the 500,000+ customer emails
3. Visit Channels -> Emails
4. Experience loading times of 10+ minutes

Steps to test this PR:

1. Import 500,000+ customer emails
2. Add at least one email and assign it to a segment with the 500,000+ customer emails
3. Visit Channels -> Emails
4. Experience loading times of < 10 seconds

[javjim]

working properly thanks!!