How to split 2 numbers into the equal ranges in PostgreSQL? - sql

Question

In PostgreSQL database I have table called layers. It looks like this:
| ID | TOTAL_SUBSCRIBERS | DENSITY |
|----|-------------------|---------|
| 1 | 34440 | |
| 2 | 41994 | |
| 3 | 102824 | |
| 4 | 19608 | |
| 5 | 1287 | |
| 6 | 4944 | |
I found max and min values of the TOTAL_SUBSCRIBERS column.
select
MIN(total_subscribers),
MAX(total_subscribers)
from
layers;
Right now I need to split the max and min into 6 range and check if each TOTAL_SUBSCRIBERS included in a certain interval. Depending on which interval is included in TOTAL_SUBSCRIBERS, I need to write the number of the interval in the DENSITY column.
For example in this table max value is 102824, min value is 1287.
RANGES:
102824 - 1287 = 101537
101537 / 6 = 16922.8333 ~ 16923
1 range: [1287-18210]
2 range: [18211-35133]
3 range: [35134-52056]
4 range: [52057-68979]
5 range: [68980-85902]
6 range: [85903-102825]
FINAL RESULT:
| ID | TOTAL_SUBSCRIBERS | DENSITY |
|----|-------------------|---------|
| 1 | 34440 | 3 | < 34440 in 3 range
| 2 | 41994 | 3 | < 41994 in 3 range
| 3 | 102824 | 6 | < 102824 in 6 range
| 4 | 19608 | 2 | < 19608 in 2 range
| 5 | 1287 | 1 | < 1287 in 1 range
| 6 | 4944 | 1 | < 4944 in 1 range

Answer 1

In a CTE calculate the min and max of TOTAL_SUBSCRIBERS and also the length of each interval and then cross join to the table to make the calculation:
with cte as (
select
min(TOTAL_SUBSCRIBERS) minsub,
((max(TOTAL_SUBSCRIBERS) - min(TOTAL_SUBSCRIBERS)) + 1) / 6 dist
from layers
)
select l.*,
(l.TOTAL_SUBSCRIBERS - c.minsub) / c.dist + 1 DENSITY
from layers l cross join (select * from cte) c
See the demo.
Results:
| id | total_subscribers | density |
| --- | ----------------- | ------- |
| 1 | 34440 | 2 |
| 2 | 41994 | 3 |
| 3 | 102824 | 6 |
| 4 | 19608 | 2 |
| 5 | 1287 | 1 |
| 6 | 4944 | 1 |
In your expected results the row with id = 1 should have DENSITY = 2, right?
Also your ranges should be:
1 range: [1287-18209]
2 range: [18210-35132]
3 range: [35133-52055]
4 range: [52056-68978]
5 range: [68979-85901]
6 range: [85902-102824]
so they are equally distrbuted.

Answer 2

I believe a combination of a generate_series and a int4range containment operation might be what you're looking for. The following code is tested on PostgreSQL 11 - see db fiddle, but should also work with 9.4+.
Sample data
CREATE TEMPORARY TABLE layers (id SERIAL, total_subscribers INT);
INSERT INTO layers (total_subscribers)
VALUES (34440),(41994),(102824),(19608),(1287),(4944);
Query
SELECT id, total_subscribers,range_id AS density,range_min,range_max
FROM layers,(
WITH j AS (
SELECT min(total_subscribers) AS min_value,
max(total_subscribers) AS max_value,
(max(total_subscribers)-min(total_subscribers))/count(*) AS var
FROM layers)
SELECT
generate_series(1,(SELECT count(*) FROM layers)) AS range_id,
generate_series(j.min_value, j.max_value-var, var)::INT AS range_min,
generate_series(j.min_value+var, j.max_value+var, var+min_value)::INT AS range_max
FROM j) j
WHERE layers.total_subscribers <# int4range(j.range_min, range_max)
ORDER BY id;
id | total_subscribers | density | range_min | range_max
----+-------------------+---------+-----------+-----------
1 | 34440 | 2 | 18209 | 36418
2 | 41994 | 3 | 35131 | 54627
3 | 102824 | 6 | 85897 | 109254
4 | 19608 | 2 | 18209 | 36418
5 | 1287 | 1 | 1287 | 18209
6 | 4944 | 1 | 1287 | 18209
(6 Zeilen)
Further reading: Common Table Expressions (CTE)

Answer 3

select the Min value and divide the value of Max - Min by six as Range from layers as table b.
select min(TOTAL_SUBSCRIBERS) as M,
(max(TOTAL_SUBSCRIBERS)-min(TOTAL_SUBSCRIBERS))/6 as R from layers b
Then select all data from layers and using TOTAL_SUBSCRIBERS to minus Min(TOTAL_SUBSCRIBERS) and divide by Range plus 1 then you can know which range(1-6) the TOTAL_SUBSCRIBERS is.
select a.*,((a.TOTAL_SUBSCRIBERS-b.M)/b.R)+1 as DENSITY from(
select layers.ID ,layers.TOTAL_SUBSCRIBERS from layers )a,
(select min(TOTAL_SUBSCRIBERS) as M,
(max(TOTAL_SUBSCRIBERS)-min(TOTAL_SUBSCRIBERS))/6 as R from layers) b