Generate date range data group based on data by dates - sql

Question

I have a list of date's data in daily basis below:
| Daytime | Item | Category| Value |
| -------- |------|------- |-------|
| 01.01.2022|A |1 |500 |
| 02.01.2022|A |1 |500 |
| 03.01.2022|A |1 |80000 |
| 04.01.2022|A |1 |500 |
| 05.01.2022|A |1 |500 |
| 01.01.2022|A |2 |600 |
| 02.01.2022|A |2 |600 |
| 03.01.2022|A |2 |600 |
| 04.01.2022|A |2 |600 |
| 05.01.2022|A |2 |600 |
| 01.01.2022|C |1 |600 |
| 02.01.2022|C |1 |600 |
| 03.01.2022|C |1 |600 |
| 04.01.2022|C |1 |600 |
| 05.01.2022|C |1 |600 |
How can i transform the data into this form?
| FromDate | ToDate | Item |Category| Value |
| --------- |--------- |------|------ |-------|
| 01.01.2022| 02.01.2022|A |1 |500 |
| 03.01.2022| 03.01.2022|A |1 |80000 |
| 04.01.2022| 05.01.2022|A |1 |500 |
| 01.01.2022| 05.01.2022|A |2 |600 |
| 01.01.2022| 05.01.2022|C |1 |600 |
I want to group the value (by item and category too) only if they are same for consecutive dates, please help, thank you!
Date format in DD.MM.YYYY and daytime's datatype is Date.
Following script for questions:
(SELECT to_date('01/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 1 Category, 500 Value FROM dual UNION ALL
SELECT to_date('02/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 1 Category, 500 Value FROM dual UNION ALL
SELECT to_date('03/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 1 Category, 80000 Value FROM dual UNION ALL
SELECT to_date('04/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 1 Category, 500 Value FROM dual UNION ALL
SELECT to_date('05/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 1 Category, 500 Value FROM dual UNION ALL
SELECT to_date('01/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 2 Category, 600 Value FROM dual UNION ALL
SELECT to_date('02/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 2 Category, 600 Value FROM dual UNION ALL
SELECT to_date('03/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 2 Category, 600 Value FROM dual UNION ALL
SELECT to_date('04/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 2 Category, 600 Value FROM dual UNION ALL
SELECT to_date('05/01/2022', 'dd/mm/yyyy') daytime, 'A' Item, 2 Category, 600 Value FROM dual UNION ALL
SELECT to_date('01/01/2022', 'dd/mm/yyyy') daytime, 'C' Item, 1 Category, 600 Value FROM dual UNION ALL
SELECT to_date('02/01/2022', 'dd/mm/yyyy') daytime, 'C' Item, 1 Category, 600 Value FROM dual UNION ALL
SELECT to_date('03/01/2022', 'dd/mm/yyyy') daytime, 'C' Item, 1 Category, 600 Value FROM dual UNION ALL
SELECT to_date('04/01/2022', 'dd/mm/yyyy') daytime, 'C' Item, 1 Category, 600 Value FROM dual UNION ALL
SELECT to_date('05/01/2022', 'dd/mm/yyyy') daytime, 'C' Item, 1 Category, 600 Value FROM dual)

Answer 1

You can use common table expression (cte) technique for that purpose.
with YourSample ( Daytime, Item, Category, Value) as (
select to_date('01.01.2022', 'DD.MM.YYYY'), 'A', 1, 500 from dual union all
select to_date('02.01.2022', 'DD.MM.YYYY'), 'A', 1, 500 from dual union all
select to_date('03.01.2022', 'DD.MM.YYYY'), 'A', 1, 80000 from dual union all
select to_date('04.01.2022', 'DD.MM.YYYY'), 'A', 1, 500 from dual union all
select to_date('05.01.2022', 'DD.MM.YYYY'), 'A', 1, 500 from dual union all
select to_date('01.01.2022', 'DD.MM.YYYY'), 'A', 2, 600 from dual union all
select to_date('02.01.2022', 'DD.MM.YYYY'), 'A', 2, 600 from dual union all
select to_date('03.01.2022', 'DD.MM.YYYY'), 'A', 2, 600 from dual union all
select to_date('04.01.2022', 'DD.MM.YYYY'), 'A', 2, 600 from dual union all
select to_date('05.01.2022', 'DD.MM.YYYY'), 'A', 2, 600 from dual union all
select to_date('01.01.2022', 'DD.MM.YYYY'), 'C', 1, 600 from dual union all
select to_date('02.01.2022', 'DD.MM.YYYY'), 'C', 1, 600 from dual union all
select to_date('03.01.2022', 'DD.MM.YYYY'), 'C', 1, 600 from dual union all
select to_date('04.01.2022', 'DD.MM.YYYY'), 'C', 1, 600 from dual union all
select to_date('05.01.2022', 'DD.MM.YYYY'), 'C', 1, 600 from dual
)
, YourSampleRanked (Daytime, Item, Category, Value, rnb) as (
select Daytime, Item, Category, Value
, row_number()over(PARTITION BY ITEM, CATEGORY ORDER BY DAYTIME) rnb
from YourSample
)
, cte (Daytime, Item, Category, Value, rnb, grp) as (
select Daytime, Item, Category, Value, rnb, 1 grp
from YourSampleRanked
where rnb = 1
union all
select t.Daytime, t.Item, t.Category, t.Value, t.rnb
, decode( t.Value, c.Value, c.grp, c.grp + 1 ) grp
from YourSampleRanked t
join cte c
on ( c.Category = t.Category and c.Item = t.Item and t.rnb = c.rnb + 1 )
)
select min(DAYTIME) FromDate, max(DAYTIME) ToDate, ITEM, CATEGORY, min(Value) Value
from cte
GROUP BY GRP, ITEM, CATEGORY
order by ITEM, CATEGORY, FromDate
;
demo on fiddle<>db
You can also use the MATCH_RECOGNIZE clause for the same purpose if you are running Oracle 12c and later.
select FromDate, toDate, ITEM, CATEGORY, VALUE
from YourSample
MATCH_RECOGNIZE (
PARTITION BY ITEM, CATEGORY
ORDER BY DAYTIME
MEASURES first(STRT.VALUE) as VALUE,
first(STRT.DAYTIME) as FromDate,
nvl(last(SAME.DAYTIME), first(STRT.DAYTIME)) as toDate
ONE ROW PER MATCH
PATTERN (STRT Same*)
DEFINE
Same AS VALUE = PREV(VALUE)
) MR
ORDER BY ITEM, CATEGORY, FromDate, toDate
;
demo2 on fiddle

Answer 2

From Oracle 12, you can use MATCH_RECOGNIZE to perform row-by-row processing:
SELECT *
FROM table_name
MATCH_RECOGNIZE (
PARTITION BY item, category
ORDER BY daytime
MEASURES
FIRST(daytime) AS from_date,
LAST(daytime) AS to_date,
FIRST(value) AS value
ONE ROW PER MATCH
PATTERN (same_value+)
DEFINE
same_value AS FIRST(value) = value
)
Which, for the sample data:
CREATE TABLE table_name (daytime, item, category, value) AS
SELECT DATE '2022-01-01', 'A', 1, 500 FROM DUAL UNION ALL
SELECT DATE '2022-01-02', 'A', 1, 500 FROM DUAL UNION ALL
SELECT DATE '2022-01-03', 'A', 1, 80000 FROM DUAL UNION ALL
SELECT DATE '2022-01-04', 'A', 1, 500 FROM DUAL UNION ALL
SELECT DATE '2022-01-05', 'A', 1, 500 FROM DUAL UNION ALL
SELECT DATE '2022-01-01', 'A', 2, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-02', 'A', 2, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-03', 'A', 2, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-04', 'A', 2, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-05', 'A', 2, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-01', 'C', 1, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-02', 'C', 1, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-03', 'C', 1, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-04', 'C', 1, 600 FROM DUAL UNION ALL
SELECT DATE '2022-01-05', 'C', 1, 600 FROM DUAL
Outputs:
ITEM
CATEGORY
FROM_DATE
TO_DATE
VALUE
A
1
2022-01-01 00:00:00
2022-01-02 00:00:00
500
A
1
2022-01-03 00:00:00
2022-01-03 00:00:00
80000
A
1
2022-01-04 00:00:00
2022-01-05 00:00:00
500
A
2
2022-01-01 00:00:00
2022-01-05 00:00:00
600
C
1
2022-01-01 00:00:00
2022-01-05 00:00:00
600
db<>fiddle here

Answer 3

This is a job for a GROUP BY using TRUNC(daytime, 'MM'). TRUNC(), when used with dates, truncates them to the beginning of a calendar / clock period.
SELECT TRUNC(Daytime, 'MM') FromDate,
ADD_MONTHS(TRUNC(Daytime, 'MM'), 1) ToDate,
Item, Category,
SUM(Value) Value
FROM my_table
GROUP BY TRUNC(Daytime, 'MM'), Item, Category
Or alternatively you can avoid those arcane Oracle date format specifiers like 'MM' and go with LAST_DAY().
SELECT ADD_MONTHS(LAST_DAY(Daytime) + 1, -1) FromDate,
LAST_DAY(Daytime) + 1 ToDate,
Item, Category,
SUM(Value) Value
FROM my_table
GROUP BY LAST_DAY(Daytime), Item, Category