Question

I have got a problem with an SQLite select.

There are the tables “kategorie”,”lv”,”majetok”,”parcela”.

Each person has information about what he owns in the table “majetok”. The sample record:

id      idpodfk idlvfk  podiel  podield cislozlv datum1         datum2
1       31      1       1/10000 0,0001  789      16. 9. 2012    16. 9. 2012

idpodfk-fk of person id
idlvfk-fk of land record id

There is also table LV (table of land records), sample record:

idlv    lvnazov katuzemie
1       1830    Plavecký Mikuláš

Also I use the table “parcela”, I store in there information about parcels in the land record

pnazov  rozloha idlvfk  idkategoriafk typ
5432    692312  1       1             C

And table kategorie

idkategoria  meno 
1            Lesné pozemky

---

What I need:
a foreach record in this select

SELECT   SUM(podield) AS sum1, idlvfk
FROM     majetok a
WHERE    idpodfk = 1
GROUP BY idlvfk

I need to select this

SELECT     SUM(par.rozloha)*sum1 AS m2, kat.meno --sum1 from statement above
FROM       Parcela par 
INNER JOIN kategorie kat ON par.IDkategoriaFK = kat.IDkategoria
WHERE      par.IDLVFK = 1 --IDLVFK from statement above
GROUP BY   kat.meno

The (second’s) query output should look like this

m2     kat.meno
123.23 Lesne pozemky    --FOR FIRST IDLVFK
324.52 Ostatne pozemky  --FOR FIRST IDLVFK   
235    Pasienky         --FOR FIRST IDLVFK 
144.23 Lesne pozemky    --FOR NEXT IDLVFK
543.52 Ostatne pozemky  --FOR NEXT IDLVFK
756    Pasienky         --FOR NEXT IDLVFK
.
.
.

And then I need to group by kat.meno so the final output should look like this:

m2     kat.meno
267,46 Lesne pozemky
868,04 Ostatne pozemky
991    Pasienky
.                         --other kat.meno if there is
.                         --other kat.meno if there is

Is this possible in one query or do I need to process it on a front-end?

Tables and query background:

There are some parcels in land record (every parcel has its own category, eg forest parcel etc). Person owns a part of a land record (saved in table majetok, eg: 1/1000). That mean he owns 1/1000 from each parcel in that land record. I want to select the area (column “rozloha” in parcela table) that he owns for each category (SUM(par.rozloha)*sum1 AS m2). So we must count sum of all parcels in that category on that land record and multiply with part that owns that person. But he can own 1/1000 on one land record and 3/3456 on the second land record (another land record=another parcels) so we have to do for each land record and that summarize you can see in last code.

Answer 1

A join over all four tables would not have the correct number of records for the sums.

What you can do is to use subqueries in place of views:

SELECT SUM(step2.area_sum),
       step2.meno
FROM (SELECT SUM(rozloha) * sum(podield_sum) AS area_sum,
             par.idlvfk,
             meno
      FROM parcela par
      JOIN kategorie kat ON par.idkategoriafk = kat.idkategoria
      JOIN (SELECT SUM(podield) AS podield_sum,
                   idlvfk
            FROM majetok
            GROUP BY idlvfk) AS ps ON ps.idlvfk = par.idlvfk
      GROUP BY idlvfk,
               meno) AS step2
GROUP BY step2.meno

I’ve tested with example data on SQL Fiddle: http://sqlfiddle.com/#!2/626e6/9