Postgres：获取计数并选择不同的

Question

I'm hitting a bit of a wall here with a pretty straightforward task (I think).

I have a table that looks something like this:

SELECT
        title, lat, lng, geocoder_identifier
    FROM entries
    WHERE geocoder_identifier IS NOT NULL;

If a user selects a place at the same lat/lng to another then the geocoder_identifier field is the same for both entries. What I need to do is group together any rows with the same geocoder_identifier, get a count of how many are in this group and then return the details for the first row in that group and discard the others.

The closest I've come to getting this done so far has been:

SELECT DISTINCT
        (geocoder_identifier), title, lat, lng, entry_count
    FROM (
        SELECT
                title,description, lat, lng, geocoder_identifier, COUNT(*) OVER (PARTITION BY geocoder_identifier) AS entry_count
            FROM entries
            WHERE geocoder_identifier IS NOT NULL) AS e
    ORDER BY e.entry_count DESC;

This produces the following dataset for me:

The problem here is that I cant really figure out how to make the results distinct by their geocoder_identifiers.

Answer 1

The simplest way would be:

SELECT min(title), min(lat), min(lng), geocoder_identifier, count(*)
FROM entries
WHERE geocoder_identifier IS NOT NULL
group by geocoder_identifier
order by 5 desc

Note that this would return the smallest values for each of title, lat and lng for each given geocoder_identifier, not necessarily the first (although lat and lng should be the same throughout each geocoder_identifier) - it depends on how you want to determine the first title.