Databases
StackOverflow2013
Postsdbo
POcounting entries, pooling over columns and returning boolean values
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POcounting entries, pooling over columns and returning boolean values
    primarykey
    data
    text
    <p>(Feel free to modify the title accordingly)</p> <p>Hi, I am using Postgres 9.3 on MacOSX. I am also new to any kind of SQL so bare with me. I am trying to create a query function (code and example below), which counts distinct entries (<code>individuals</code>) by pooling over certain columns by using <code>SELECT DISTINCT ON</code>. This works fine. The problem arrises when I in the same function add a <code>CASE</code> statement for checking if each rows of yet another column returns <code>NULL</code> or NOT.</p> <p>So, I have a table looking like this but with thousands of entries. <a href="http://i1326.photobucket.com/albums/u642/Awful-Dodger/ScreenShot2013-11-19at93404PM_zps12da3754.png" rel="nofollow">Here is a subset</a> where you can see data for a particular species and method. The geographical data are of different resolution, but here I am pooling over the columns <code>species, method, location, larger_region</code> (in example, pooling over <code>method</code> is redundant) See code and output below.</p> <pre><code>--CODE CREATE TABLE test_tbl (accn varchar(10) PRIMARY KEY, method TEXT, species VARCHAR(30), lat NUMERIC, lon NUMERIC, location TEXT, larger_region TEXT, coords BOOLEAN DEFAULT TRUE); FETCH FORWARD 1 FROM test_tbl; accn | method | species | lat | lon | location | larger_region | coords ---------+----------+-------------------+----------+----------+----------------------------+---------------+--------- EU336944 | Isolate | Dysidea granulosa | 10.07704 | 73.63238 | Lakshadweep Islands, India | Indian EEZ | t CREATE TABLE test_view_tbl ( species VARCHAR(30) NOT NULL, no_ind BIGINT NOT NULL, method TEXT NOT NULL, location TEXT NOT NULL, larger_region TEXT NOT NULL ); </code></pre> <p>Below function produces the correct output.</p> <pre><code>CREATE FUNCTION test_view() RETURNS SETOF test_view_tbl AS $$ SELECT DISTINCT ON (species, method, location, larger_region) species, count(species), method, location, larger_region FROM test_tbl GROUP BY species, method, location, larger_region ORDER BY species; $$ LANGUAGE sql SELECT * FROM test_view(); species | no_ind | method | location | larger_region ------------------+--------+---------+--------------------------------------+--------------- Dysidea granulosa | 6 | Isolate | Kavaratti Island, Lakshadweep, India | Indian EEZ Dysidea granulosa | 13 | Isolate | Lakshadweep Islands, India | Indian EEZ </code></pre> <p>Below I modify <code>test_view_tbl</code> by adding a <code>boolean</code> column <code>coords</code> and modify the code correspondingly. This checks if column <code>lat</code> is <code>NULL</code> or not, i.e. whether each entry has coordinates. However, this also screws up the above wanted result by NOT returning certain entries. </p> <pre><code>ALTER TABLE test_view_tbl ADD COLUMN coords BOOLEAN DEFAULT TRUE; CREATE FUNCTION test_view() RETURNS SETOF test_view_tbl AS $$ SELECT DISTINCT ON (species, method, location, larger_region) species, count(species), method, location, larger_region, CASE WHEN lat IS NULL THEN coords IS FALSE ELSE TRUE END FROM test_tbl GROUP BY species, method, location, larger_region, lat, coords ORDER BY species; $$ LANGUAGE sql SELECT * FROM test_view(); species | no_ind | method | location | larger_region | coords -------------------+--------+---------+--------------------------------------+---------------+-------- Dysidea granulosa | 6 | Isolate | Kavaratti Island, Lakshadweep, India | Indian EEZ | t Dysidea granulosa | 3 | Isolate | Lakshadweep Islands, India | Indian EEZ | t </code></pre> <p>So, each part of the code works, but apparently not together since it returns <code>NULL</code> on 10 entries. It seems that it only counts each distinct coordinates (<code>lat</code>) within <code>location</code> <code>Lakshadweep Islands, India</code> I have tried to fix it but without success. I am hoping for some pointers, thanks </p>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. USjO.
    1. USjO.
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. COI figure you want to aggregate on {species+location fields} How do you want the boolean `coords` to be _pooled_ in the aggregate ?
      singulars
      1. POcounting entries, pooling over columns and returning boolean values
      1. USjoop
    2. COFor the fields that are **not part** of the `GROUP BY` clause (or the `DISTINCT ON ...`, which is similar), the only thing you can do with them is aggregate them: take their MAX() or MIN() or there count(). (`distinct on` is a _bit_ different). What do you want to happen wrt the booleans, count the number (or proportion) of True observations in the aggregate ? Please Note: you can only make _one_ value (per aggregated group) appear in the result!
      singulars
      1. POcounting entries, pooling over columns and returning boolean values
      1. USjoop
    3. CO@joop: Thanks for commenting. **EDITED** Yes, you are right. I wanna pool over those fields but not over coords... even if coords are the only thing different (f / t) when the pooled fields are the same, as in the above example, I want the corresponding entries to be accounted for, thus the potential to have the same lines but with different boolean `(f / t)` values. Sorry for being unclear enough, it's too late at my end. It's mostly so I can get a feel of how many `species-location` fields I have coordinates for, for further analyses.
      singulars
      1. POcounting entries, pooling over columns and returning boolean values
      1. USjO.
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload