I have this query, which works:
SELECT TO_CHAR(last_date_called,'HH24'), count(*)
FROM log_table
GROUP BY TO_CHAR(last_date_called,'HH24');
But, in some cases there are not 24 hours worth of data. What I want to do, is always generate 24 rows, and if there is nothing for that hour, return 0. So, results may look like this:
00 10
01 25
02 33
03 0
04 55
05 0
06 23
And so on........
You'll need a row generator to create all hours in a day, and then outer join it to your "real" table. Something like this (see comments within code):
SQL> with
2 hours as
3 -- row generator, to create all hours in a day
4 (select lpad(level - 1, 2, '0') hour
5 from dual
6 connect by level <= 24
7 ),
8 log_table (last_date_called) as
9 -- sample data, just to return "something"
10 (select to_date('08.07.2021 13:32', 'dd.mm.yyyy hh24:mi') from dual union all
11 select to_date('16.02.2021 08:20', 'dd.mm.yyyy hh24:mi') from dual
12 )
13 -- final query
14 select h.hour,
15 count(l.last_date_called) cnt
16 from hours h left join log_table l on h.hour = to_char(l.last_date_called, 'hh24')
17 group by h.hour
18 order by h.hour;
HO CNT
-- ----------
00 0
01 0
02 0
03 0
04 0
05 0
06 0
07 0
08 1
09 0
10 0
11 0
12 0
13 1
14 0
15 0
16 0
17 0
18 0
19 0
20 0
21 0
22 0
23 0
24 rows selected.
SQL>
Related
The problem is quite simple to understand but solving it was not as easy as it sounded at first.
Let's assume the following, an image that is 8*4, normal order is easy, you return the pixel index:
// 00 01 02 03 04 05 06 07
// 08 09 10 11 12 13 14 15
// 16 17 18 19 20 21 22 23
// 24 25 26 27 28 29 30 31
Now suppose you want to swizzle rows like so:
// 00 01 02 03 04 05 06 07
// 16 17 18 19 20 21 22 23
// 08 09 10 11 12 13 14 15
// 24 25 26 27 28 29 30 31
I solved it, not without trouble to be honest, with the following formula:
index / 8 % 2 * 16 + index / 16 * 8 + index % 8
Isn't there a simpler formula to get the same result?
Assuming / and % return the quotient and remainder in the Euclidean division:
The classic ordering can be obtained as:
row = n / 8
col = n % 8
And the swizzled ordering can be obtained as:
col = n % 8
old_row = n / 8
new_row = 2 * (old_row / 2) + (1 - (old_row % 2))
Explanation:
2 * (old_row / 2) groups the rows two by two;
(1 - (old_row % 2)) swaps row 0 and row 1 of each group.
I have the below table and I am trying to write an Oracle select query that would result in the second table.
I understand I have to use Pivot but I can't figure at all how to do.
MSGID
KEY
COLVALUE
15
height
18
15
length
19
15
width
20
15
notImportant
xxx
16
height
21
16
length
22
16
width
23
16
notImportant
xxx
17
height
24
17
length
25
17
width
26
17
notImportant
xx
Desired result:
MsgID
height
length
width
15
18
19
20
16
21
22
23
17
24
25
26
Tried the below code but without success....
select MSGID, HEIGHT, LENGTH, WIDTH
from (select MSGID, KEY, COLVALUE
from table )
PIVOT
(
max(COLVALUE)
FOR KEY IN ('HEIGHT','LENGTH','WIDTH')
)
Do you have any tips?
Here's one option:
SQL> select msgid,
2 max(case when key = 'height' then colvalue end) height,
3 max(case when key = 'length' then colvalue end) length,
4 max(case when key = 'width' then colvalue end) width
5 from test
6 group by msgid
7 order by msgid;
MSGID HEIGHT LENGTH WIDTH
---------- ---------- ---------- ----------
15 18 19 20
16 21 22 23
17 24 25 26
SQL>
Or, with pivot:
SQL> select *
2 from test
3 pivot
4 (max(COLVALUE)
5 FOR KEY IN ('height','length','width')
6 )
7 order by 1;
MSGID 'he 'le 'wi
---------- --- --- ---
15 18 19 20
16 21 22 23
17 24 25 26
SQL>
I have this table:
Year
Month
Agency
Value
2019
9
1
233
2019
9
4
132
2019
8
3
342
2020
3
2
321
2020
3
4
34
2020
5
2
56
2020
5
4
221
2020
5
1
117
2018
12
2
112
2018
12
2
411
2020
4
3
241
2020
4
2
155
I'd like to set a new measure/column where last month from last year is 1, and 0 in another cases:
Year
Month
Agency
Value
Filter
2019
9
1
233
0
2019
9
4
132
0
2019
8
3
342
0
2020
3
2
321
0
2020
3
4
34
0
2020
5
2
56
1
2020
5
4
221
1
2020
5
1
117
1
2018
12
2
112
0
2018
12
2
411
0
2020
4
3
241
0
2020
4
2
155
0
I've been able to "copy" a new table with values from Month=5 and Year=2020 ("the lastest from the lastest"):
TableData - Last Charge =
var table = FILTER(
TableData,
AND(
MAX('TableData '[Year])='TableData '[Year],
MAX('TableData '[Month])='TableData '[Month]
)
)
return SUMMARIZE(table , TableData [Year], TableData [Month], TableData [Agency], TableData [Value])
However, my intention is don't create new tables and use measures/columns tu use it like FILTER when I create a graphic.
Thanks a lot, and sorry for my poor english.
I solved it with this measure:
Measure =
VAR a =
MAX ( 'Table'[Year] )
VAR b =
MAX ( 'Table'[Months] )
VAR c =
MAXX ( ALL ( 'Table' ), [Year] )
VAR d =
MAXX ( FILTER ( ALL ( 'Table' ), [Year] = c ), [Months] )
RETURN
IF ( a * 100 + b = c * 100 + d, 1, 0 )
I have a problem trying to group some records and trying to count the number of records returned by a query is adding an example:
07 COMERCIO 92
15 SERVICIOS OTROS 41
01 AGRICULTURA, GANADERIA Y SILVICULTURA 141
04 INDUSTRIA MANUFACTURERA 28
10 BANCA Y FINANZAS 5
12 ADMINISTRACION PUBLICA 16
03 MINERIA 3
16 HOGAR 2
08 HOTELES Y RESTAURANTES 37
11 EMPRESARIAL 21
14 SOCIAL Y SALUD 4
06 CONSTRUCCIÓN 3
09 TRANSPORTE 30
13 EDUCACION 10
This is query:
SELECT
AGRUP.VC_CODDET AS CHR_SECECO,
AGRUP.VC_NOMDET AS VC_SECECO,
0 AS INT_NROPRESTAMO,
COUNT(*) INTO_BENEFICIARIOS,
0 AS DEC_SALCON
FROM TB_JSI_PRESTAMO_DETALLE PREDET
INNER JOIN TB_JSI_PRESTAMO PRE ON
PRE.INT_IDPRESTAMO=PREDET.INT_IDPRESTAMO
INNER JOIN TB_JSI_BENEFICIARIO_IFI BENIFI ON
BENIFI.INT_IDBENEIFI=PRE.INT_IDBENEIFI
INNER JOIN TB_JSI_OPERACION OPE ON PRE.INT_IDOPE = OPE.INT_IDOPE AND
TRUNC(TO_DATE(OPE.DT_FECVEN))>TRUNC(TO_DATE('30/09/2018', 'DD/MM/YY'))
INNER JOIN TB_JSI_CIIU CIIU ON PRE.INT_IDACT = CIIU.INT_IDACT AND
PRE.INT_IDSEC=CIIU.INT_IDCLAS
INNER JOIN TB_JSI_TABLA_DET SECECO ON SECECO.INT_IDDET=CIIU.INT_IDCLAS
INNER JOIN TB_JSI_AGRUPA_SECTOR AGRSEC ON SECECO.INT_IDDET =
AGRSEC.INT_IDSEC
INNER JOIN TB_JSI_TABLA_DET AGRUP ON AGRSEC.INT_IDAGRU = AGRUP.INT_IDDET
INNER JOIN TB_JSI_TABLA_DET MON ON OPE.INT_IDMON = MON.INT_IDDET
INNER JOIN TB_JSI_IFI IFI ON OPE.INT_IDIFI = IFI.INT_IDIFI
WHERE TRUNC(PREDET.DTE_FECPRO) = (
SELECT
TRUNC(DTE_FECPRO)
FROM (SELECT
DTE_FECPRO
FROM TB_JSI_PRESTAMO_DETALLE
WHERE DTE_FECPRO<=TO_DATE('30/09/2018','DD/MM/YY')
AND DEC_SALDOL>0
ORDER BY DTE_FECPRO DESC)
WHERE ROWNUM = 1
)
AND (NULL IS NULL OR OPE.INT_IDTIPPRO =2)
AND (NULL IS NULL OR OPE.INT_IDMON = 364)
AND (NULL IS NULL OR OPE.INT_IDIFI=72)
GROUP BY AGRUP.VC_CODDET, AGRUP.VC_NOMDET
and should return this result
07 COMERCIO 92
15 SERVICIOS OTROS 41
01 AGRICULTURA, GANADERIA Y SILVICULTURA 141
04 INDUSTRIA MANUFACTURERA 28
10 BANCA Y FINANZAS 4
12 ADMINISTRACION PUBLICA 16
03 MINERIA 3
16 HOGAR 2
08 HOTELES Y RESTAURANTES 37
11 EMPRESARIAL 21
14 SOCIAL Y SALUD 4
06 CONSTRUCCIÓN 3
09 TRANSPORTE 30
13 EDUCACION 10
since I have these records and when counting I should return 4 and not 5
269516 10 BANCA Y FINANZAS 1
269558 10 BANCA Y FINANZAS 1
269592 10 BANCA Y FINANZAS 2
269611 10 BANCA Y FINANZAS 1
#Roller, I have analyzed the result and it is now clear that one of the tables used in the query has two records with the matching condition and that is why it is returning 5 and not 4.
You just need to identify the table and if you are not able to, then use the following trick.
First of all, identify the table, basis of which you are saying that there must be 4 rows and then use that table's PK in the count with distinct as follows:
COUNT(DISTINCT MY_PK)
I have this task and i can't figure out how to do it.
I need to find persons age in days, there are given birth and death dates, there's data file:
8
Albertas Einšteinas 1879 03 14 1955 04 18
Balys Sruoga 1896 02 02 1947 10 16
Antanas Vienuolis 1882 04 07 1957 08 17
Ernestas Rezerfordas 1871 08 30 1937 10 17
Nilsas Boras 1885 10 07 1962 11 18
Nežiniukas Pirmasis 8 05 24 8 05 25
Nežiniukas Antrasis 888 05 25 888 05 25
Nežiniukas Trečiasis 1 01 01 125 01 01
and there's how result file should look like:
1879 3 14 1955 4 18 27775
1896 2 2 1947 10 16 18871
1882 4 7 1957 8 17 27507
1871 8 30 1937 10 17 24138
1885 10 7 1962 11 18 28147
8 5 24 8 5 25 1
888 5 25 888 5 25 0
1 1 1 125 1 1 45260
Few things to notice: all februarys have 28 days.
My function for calculating age:
function AmziusFunc(Mas : TZmogus) : longint;
var amzius, max : longint;
begin
max := 125 * 365;
amzius := (Mas.mirY - Mas.gimY) * 365 + (Mas.mirM - Mas.gimM) * 31 +
(Mas.mirD - Mas.gimD);
if ( amzius >= max ) then amzius := 0;
AmziusFunc := amzius;
end;
What should i change there? Thanks.
function AmziusFunc(Mas : TZmogus) : longint;
var amzius, max : longint;
begin
max := 125 * 365;
amzius := (Mas.mirY - Mas.gimY) * 365 + (Mas.mirM - Mas.gimM) * 31 +
(Mas.mirD - Mas.gimD);
if ( amzius >= max ) then amzius := 0;
AmziusFunc := amzius;
end;