Update column using a calculation based on a query

Question

Hi,
I need to update an empty column I have created in my table. I have created a query that returns the variables I will need in the calculation but I'm not sure how to link it.
This query works fine.
Select z.seq_no
, z.Day_X
, z.Day_Y
, x.Night_X
, x.Night_Y
from
(select seq_no, ga.centre_x Day_X, ga.centre_y Day_Y
from MICHAELC_Nov_01_final ff, rbn_gis_area ga
where ff.day_addr_code = ga.sp_code) z
,(select seq_no, ga.centre_x Night_X, ga.centre_y Night_Y
from MICHAELC_Nov_01_final ff, rbn_gis_area ga
where ff.risk_addr_code = ga.sp_code) x
where z.seq_no = x.seq_noWhat I require is my column called Distance_to_work to be calculated as follows:
Distance_to_work = (100 * sqrt (power ( (Day_X - Night_X), 2) + power ( (Day_Y - Night_Y), 2)))This is what I have tried to no avail:
UPDATE MICHAELC_Nov_01_final ff
SET
Distance_to_work = (select (100 * sqrt (power ( (Day_X - Night_X), 2) + power ( (Day_Y - Night_Y), 2))))
from
(Select z.seq_no
, z.Day_X
, z.Day_Y
, x.Night_X
, x.Night_Y
from
(select seq_no, ga.centre_x Day_X, ga.centre_y Day_Y
from MICHAELC_Nov_01_final ff, rbn_gis_area ga
where ff.day_addr_code = ga.sp_code) z
,(select seq_no, ga.centre_x Night_X, ga.centre_y Night_Y
from MICHAELC_Nov_01_final ff, rbn_gis_area ga
where ff.risk_addr_code = ga.sp_code) x
where z.seq_no = x.seq_no) w
where ff.sp_code = w.sp_codeAny assistance would be appreciated.
Banner:
Oracle Database 11g Release 11.2.0.2.0 - 64bit Production
PL/SQL Release 11.2.0.2.0 - Production
"CORE 11.2.0.2.0 Production"
TNS for Linux: Version 11.2.0.2.0 - Production
NLSRTL Version 11.2.0.2.0 - Production

Answer 1

Hi,
885178 wrote:
...  with Temp_table          
as
select 77317027 sp_code, 28.137 x_day, 28.244 x_night, -26.179 y_day, -26.092 y_night, 13.7905 distance_to_work  from dual
union all
select 57218000 sp_code, 30.897 x_day, 30.851 x_night, -30.022 y_day, -29.884 y_night, 14.54647 distance_to_work  from dual
union all
select 70406009 sp_code, 28.131 x_day, 28.192 x_night, -26.316 y_day, -25.837 y_night, 48.28685 distance_to_work  from dual
select *
from Temp_table
What is temp_table? What role does it play in this problem?
with michaelc_nov_01_final2          
as
select 77317027 sp_code, 77317029 risk_addr_code, 77303000 day_addr_code, 1274799 seq_no, 13.7905 distance_to_work   from dual
union all
select 57218000 sp_code, 57271032 risk_addr_code, 57218000 day_addr_code, 1624223 seq_no, 14.54647 distance_to_work from dual
union all
select 70406009 sp_code, 77603008 risk_addr_code, 77301006 day_addr_code, 2240402 seq_no, 48.28685 distance_to_work from dual
select *
from michaelc_nov_01_final2What is micahealc_nov_01_final2? Do you want micahealc_nov_01_final to be changed so that it is just like micahealc_nov_01_final2? Can't you just post the results you want, like this:
`  SP_CODE RISK_ADDR_CODE DAY_ADDR_CODE     SEQ_NO DISTANCE_TO_WORK
  77317027       77317029      77303000    1274799       13.7905765
  57218000       57271032      57218000    1624223       14.5464772
  70406009       77603008      77301006    2240402       48.2868512?
The data above is what michaelc_nov_01_final looks like after running the MERGE statement below. Distance_to_work isn't exactly what it is in michaelc_nov_01_finalw, but it looks like you're just rounding differently.
The rbn_gis_area is versioned. There is a column called version and the latest version = 10. This will bring back unique rows. There's no version column in the code above.
with rbn_gis_area          
as
select 77303000 sp_code, 28.137 centre_x, -26.179 centre_y    from dual
union all ...
Also note that an sp_code can be day_addr_code as seen in the above example. Seq no. is not unique. There is a column in michaelc_nov_01_final called cla_case which is unique.There's no cla_case column in the code above.
with michaelc_nov_01_final          
as
select 77317027 sp_code, 77317029 risk_addr_code, 77303000 day_addr_code, 1274799 seq_no, '' distance_to_work   from dual
union all ...The following assumes that seq_no is unique:
MERGE INTO       michaelc_nov_01_final          dst
USING (
        SELECT       ff.seq_no
     ,       MIN (d.centre_x)     AS day_x
     ,       MIN (d.centre_y)     AS day_y
     ,       MIN (n.centre_x)     AS night_x
     ,       MIN (n.centre_y)     AS night_y
     FROM       michaelc_nov_01_final     ff
     JOIN       rbn_gis_area          d  ON   ff.day_addr_code   = d.sp_code
     JOIN       rbn_gis_area          n  ON     ff.risk_addr_code  = n.sp_code
     GROUP BY  ff.seq_no
)                                 src
ON      (src.seq_no     = dst.seq_no)
WHEN MATCHED THEN
UPDATE  SET     dst.distance_to_work  = 100 * SQRT ( POWER (src.day_x - src.night_x, 2)
                                             + POWER (src.day_y - src.night_y, 2)
;If cla_case unique in your real table, then use cla_case instead of seq_no above.
I used GROUP BY TO get (at most) one row in src for every row in dst. If you have a version column, you can use that instead or GROUP BY, just make sure that each row in src matches (at most) one row in dst.