Retrun separate lines with duplicate values

Hello

I have the columns id, addr1, addr2, city and State of coding. I need to know all the coding IDS that have the same addresses and at the same time filter out the same side of the id of the request.

For example:

acctid addr1, addr2 City State

WEF 1 101 101 sd

1 101 wef sd

WEF 2 101 101 sd

WEF 3 101 101 sd

DC 4 102 102 homeless

From the above data, I want to get filter tier 4 and get lines 1, 2 and 3. How to achieve with a sql query

Thank you

Arch

Wrong forum!

This forum is ONLY for sql developer questions.

Please mark the thread ANSWERED and repost it in Sql and PL/SQL forum.

SQL and PL/SQL

Tags: Database

Similar Questions

Updated several lines with different values
Hello!
I have a problem. I need to update more than 1000 lines with different values. How can I do?
For exsample I have table:
ID; color, date,
1 red
2 green
3 white

I need to update the date field.

Update table
Set date = '01.02.03'
where id = 1

Update table
Set date = '01.03.03'
where id = 2

Maybe it's how to update multiple rows in a single request?

Sorry for my bad English.
Thank you!
Hello

You can try this
```
UPDATE TABLE SET DATE = CASE
                    WHEN ID = 1 THEN TO_DATE('01-02-03','DD-MM-RR')
                    WHEN ID = 2 THEN TO_DATE('01-03-03','DD-MM-RR')
                    END
```
see you soon

VT

How to display the line empty as a line with null values
Hi all

Pls advise me if it is possible to use a single query statement to display
Empty row (i.e. not a single return line) as a line with null values.

For example,.

Select the names of names_mst whose name = "sgasfgs".

Result:
Names of
=====
< null >
Hello
If you desire to join external to double, as shown below, you still get at least a line of production
```
SELECT  nm.names
FROM            dual
LEFT OUTER JOIN names_mst   nm  ON nm.name='sgasfgs';
```

Pulling on the first line with a value of repetition
Hello
We use the Oracle 11.1.

We have a table of postal codes.

We have several lines for each zip code because we have different areas that use the same zip code.

We want to reduce the lines of a line by zip code. But they said use they want to keep the fields that differ (neighborhood of the exodus).

Then they said "Just take the first value of each."

This code reduces the postcodes of a line by postal code, but as you can see I replaced the field with the name of the field values.
```
SELECT distinct  ZIPCODE, CITY, STATE, COUNTY, AREACODE, 'CITYTYPE', CITYALIASABBREVIATION, 'CITYALIASNAME', LATITUDE, LONGITUDE, TIMEZONE, ELEVATION, COUNTYFIPS, DAYLIGHTSAVING, PREFERREDLASTLINEKEY, CLASSIFICATIONCODE,
           MULTICOUNTY, STATEFIPS, 'CITYSTATEKEY', 'CITYALIASCODE', 'PRIMARYRECORD', CITYMIXEDCASE, 'CITYALIASMIXEDCASE', STATEANSI, COUNTYANSI
    FROM   zip_code
GROUP BY   ZIPCODE, CITY, STATE, COUNTY, AREACODE, CITYTYPE, CITYALIASABBREVIATION, LATITUDE, LONGITUDE, TIMEZONE, ELEVATION, COUNTYFIPS, DAYLIGHTSAVING, PREFERREDLASTLINEKEY, CLASSIFICATIONCODE, MULTICOUNTY, STATEFIPS,
           CITYALIASCODE, PRIMARYRECORD, CITYMIXEDCASE, STATEANSI, COUNTYANSI
order by zipcode; 
```
How should I take either one record per postal code with the first value in these fields or

How should I take the first record for each zip code?

either way works for me.

Thank you
Hello

This is called a Query Top - N , and this is a way to do it:
```
WITH     got_r_num     AS
(
     SELECT     z.*
     ,     ROW_NUMBER () OVER ( PARTITION BY  zipcode
                               ORDER BY          city
                         ,                ...   -- add tie-breakers here
                       )      AS r_num
     FROM    zip_code
)
SELECT     *     -- or list all columns except r_num
FROM     got_r_num
WHERE     r_num     = 1
;
```
You can simply use MIN or MAX on the columns, as the line with the name of the minimum city do not lowest have area code or the longitude. The above method preserves the first line of each zip code intact, where the 'first': the first in order of aphabetic by the name of the city. If it is not unique, then you may want breakage to the analytical ORDER BY clause.

If you are deleting rows, you can use this in a NOT IN subquery. If you copy lines to a new table, you can use the query above in an INSERT statement.

Published by: Frank Kulash, March 18, 2011 12:40

I just see Polywog response:

Pollywog wrote:
You can use the rank or rownumber function...

If I understand the problem, you don't want to use RANK in this particular case. When there is equality, GRADE will assign 1 to all the contenders. I think the whole point of this thread is that you exactly want a #1 in each zip code, which is what ROW_NUMBER, but not RANK, guarantees.

Add filter slot table form to show the lines with null values
Hello

I have a tabular presentation where we need to add a filter with two values (selection list) so that if one of the columns in a table (column1) is zero, then I need to display all records where column1 is null.

I was wondering how do I put this in the Source of tabular forms where condition. This filter has only two values,

All the
NULL for the error

So when the user selects all, then all the records in the table are shown
If the user selects 'Value Null in error', while the records where Column1 is null will be shown

For example
```
 column1        column2
                       A
                       B
      1               C
      3               D
 
```
If the user selects, then

all four files will show up. Else, only the first and the second will be returned.

Thank you
Hi, Ryan.

Try a static LOV like this
```
STATIC2:;,--- All ---;--- All ---
```
and a Where clause, like this
```
where column1 is null or :p1_filter like '%%'
```
Jeff

Adding a UNIQUE CONSTRAINT to an existing column with duplicate values
Oracle Forums Hello, I have a table that contains rows 11 901. I tried to define the UNIQUE constraint on the column of xcode, but encountered an error: the column already contain duplicates! I rescued following SQL 'SELECT COUNT (*) FROM States;' the result is 11 901 and 'SELECT COUNT (DISTINCT (xcode)) FROM States;' the result is 11 680. So to enforce this unique key I have to remove the 221 lines. Please how can I do this?
REMOVE duplicates...

For example, as
```
delete from your_table t
where rowid <
   (
   select max(rowid)
   from your_table r
   where t.unique_column = r.unique_column);
```
Thne create the constraint.

Get line with max value
Hi all

I have a table

Amount Line_num
1 10
2 20
3 30

create table Test (number line_num, quantity number);
Insert into test values (1,10);
Insert into test values (2.20).
Insert into test values (3.30);

I need to get the line number that has a maximum amount.

Help, please.

Thank you

Anatoliy
```
select
  max(line_num) keep (dense_rank last order by amount) m
from test

M
3
```
concerning

Ridge line with default values
Hi friends I am using JDeveloper 11.1.1.3.I want to create a new line in a table .but this line must have some default values. As an attribute username, still need to show a default name. I need to write the ViewImpl method.
According to [url http://download.oracle.com/docs/cd/E17904_01/apirefs.1111/e10653/oracle/jbo/server/ViewRowImpl.html#create (oracle.jbo.AttributeList documentation)] the, really want to do this by overriding the method create of the ViewRowImpl class (that is, assuming that in the VO is the right place).

The Fusion developer [url http://download.oracle.com/docs/cd/E17904_01/web.1111/b31974/bcrules.htm#sthref794] Guide also shows you how to proceed at the level of the OS

John

generate the view in line with the values of the 'sequence' 0.36
I want my query to join inline view that has values of 0.36, what is the best way to create this inline-view?
My initial solution would be:
```
SELECT LEVEL RESULT FROM Dual CONNECT BY LEVEL < 37
union
select 0 from dual
order by 1
```
I want to write the query that returns the same values as this query, I showed.
CharlesRoos wrote:
I want my query to join inline view that has values of 0.36, what is the best way to create this inline-view?
My initial solution would be:
```
SELECT LEVEL RESULT FROM Dual CONNECT BY LEVEL < 37
union
select 0 from dual
order by 1
```
I want to write the query that returns the same values as this query, I showed.
```
SELECT (LEVEL - 1) RESULT FROM Dual CONNECT BY LEVEL <= 37 ;
```
Published by: user503699 on October 29, 2010 14:20

How to return a single LINE with Multiple value separated by a colon in a SQL query
Hello

I have a SQL, as shown.
```
select deptno
  from dept
```
I want to mofidfy this query, so that it should return me departments with colon delimeted list in a ROW.
```
10:20:30:40.......
```
Thank you
Deepak

Need to separate lines, with the help of the distinct clause
I have according to requirement

Organization of end startdate EmpID
1 O1 from January 1, 2008
2 on January 2, 2008 July 31 08 O1
3 2 January 2008 31 July 08 O1
4 2 January 2008 31 July 08 O2
5 2 January 2008 31 July 08 O2
6 2 January 2008 31 July 08 O2
7 2 January 2008 31 July 08 O3
8 2 January 2008 31 July 08 O3

My requirement is to get a sql query which report me something like that

Organization of end startdate EmpID
1 O1 from January 1, 2008
4 2 January 2008 31 July 08 O2
7 2 January 2008 31 July 08 O3

Any empid, startdate, which is fine for a given organism (IE organization should be separate)

concerning
Vikas

Published by: user634784 on January 19, 2009 04:05
In reality... N °

Sure about that?
```
SQL>  select organisation,
  2     min(startdate) keep(dense_rank first order by empid asc) startdate,
  3     min(enddate) keep(dense_rank first order by empid asc) enddate,
  4     min(empid) keep(dense_rank first order by empid asc) empid
  5     from t
  6    group by organisation
  7  /

OR STARTDATE ENDDATE        EMPID
-- --------- --------- ----------
O1 01-JAN-08                    1
O2 05-JAN-08 31-JUL-08          4
O3 02-JAN-08 31-JUL-08          7
```

Query generates duplicate values in successive lines. I can null out them?
I have had very good success (thanks to Andy) to obtain detailed responses to my questions posted, and I wonder if there is a way in a region report join query that produces lines with duplicate values, I cannot suppress (replace) printing of duplicate values in successive lines. In the event that I managed to turn the question into a mess unintelligible (one of my specialities), let me give an example:

We try institutions undergraduate that an applicant has participated in the list and display information on release dates, gpa and major (s) / according to decision-making. The problem is that there may be several major (s) / minor (s) for a given undergraduate institution, so the following is produced by the query:

Knox College hard 01/02 01/06 4,00 knitting
Knox College hard 01/02 01/06 4,00 cloth repair
Advanced University 02/06 01/08 3.75 Clothing Design
Really advanced U 02/08 01/09 4,00 sportswear
Really advanced U 02/08 01/09 4,00 basketball burlap

I want it to look like this:

Knox College hard 01/02 01/06 4,00 knitting
Tissue repair
Advanced University 02/06 01/08 3.75 Clothing Design
Really advanced U 02/08 01/09 4,00 sportswear
Burlap tennis shoe

* (Edit) Please note that the repair of fabric and lines tennis shoe repair should be positioned properly in a table, but unfortunately had space here for a reason suppresed any. *

Under the tuteage of Andy, I would say the answer is probably javascript one loop in the DOM, you are looking for the innerHTML of specific TDs in the table, but I want to confirm that, or Apex provides it a box I can check which will produce the same results? Thanks to the guy to advance and sorry for all the questions. We were charged to use Apex for our next project and learn it using it, as the training budget is non-existent this year. I love ;) unfunded mandates

Phil

Published by: Phil McDermott on August 13, 2009 09:34
Hi Phil,

JavaScript is useful, because the feature break column in the report attributes (which would be my first choice if poss).

If you need to go beyond 3 columns, I would say something in the SQL statement itself. This means that the sort would probably have to be controlled, but it is doable.

This is a pretty old thread on the subject: Re: grouping reports (non-interactive) -with an example here: [http://htmldb.oracle.com/pls/otn/f?p=33642:112] (I used a custom template for the report, but you don't need to go that far!)

This uses the features of LAG in SQL to compare the values on line for values on the line before - ability to compare these values allows you to determine which ones to show or hide.

Andy

two rows of delete with different values

Hello

I wouldn't lines with the same id and have values of E and F 10258932.

I would like the following data:

A        B  C      D     E N
-------- -- ------ ----- - -
10258927 1  103,35 0
10258929 3  284,85 89,52 E N
10258929 4  323,85 89,52 E N
10258930 5  478,80 91,53 E N
10258931 6  436,67 78,09 E N

But I have the following lines:

SQL> with my_table as
  2  (
  3  select '10258927' a, '1'  b, '103,35'  c, '0'      d,   NULL  e from dual union all    
  4  select '10258928' a, '2'  b, '0'       c, '15,19'  d,  'E'    e from dual union all
  5  select '10258929' a, '3'  b, '284,85'  c, '89,52'  d,  'E'    e from dual union all
  6  select '10258929' a, '4'  b, '323,85'  c, '89,52'  d,  'E'    e from dual union all
  7  select '10258930' a, '5'  b, '478,80'  c, '91,53'  d,  'E'    e from dual union all
  8  select '10258931' a, '6'  b, '436,67'  c, '78,09'  d,  'E'    e from dual union all
  9  select '10258932' a, '7'  b, '784,88'  c, '23,19'  d,  'E'    e from dual union all
 10  select '10258932' a, '8'  b, '100,88'  c, '11,11'  d,  'E'    e from dual union all
 11  select '10258932' a, '9'  b, '300,88'  c, '22,22'  d,  'F'    e from dual union all
 12  select '10258932' a, '10' b, '468,25'  c, '78,33'  d,  'F'    e from dual)
 13  select a, b, c, d, e, no_ds from
 14  (
 15    select a, b, c, d, e,
 16           case when c > 0 and e != 'F'
 17                then 'N'
 18           end no_ds
 19    from my_table)
 20  where no_ds = 'N' or e is null
 21  order by 1
 22  /

A        B  C      D     E N
-------- -- ------ ----- - -
10258927 1  103,35 0
10258929 3  284,85 89,52 E N
10258929 4  323,85 89,52 E N
10258930 5  478,80 91,53 E N
10258931 6  436,67 78,09 E N
10258932 7  784,88 23,19 E N
10258932 8  100,88 11,11 E N

7 Zeilen ausgewählt.

SQL>

Does anyone have an idea?

Oh, I read the title again once: 'remove double lines with different values.
If it means that he is not specifically E and F values, but only if there is more than one distinct value, then you can do the following:

with my_table as
(
select '10258927' a, '1'  b, '103,35'  c, '0'      d,   NULL  e from dual union all
select '10258928' a, '2'  b, '0'       c, '15,19'  d,  'E'    e from dual union all
select '10258929' a, '3'  b, '284,85'  c, '89,52'  d,  'E'    e from dual union all
select '10258929' a, '4'  b, '323,85'  c, '89,52'  d,  'E'    e from dual union all
select '10258930' a, '5'  b, '478,80'  c, '91,53'  d,  'E'    e from dual union all
select '10258931' a, '6'  b, '436,67'  c, '78,09'  d,  'E'    e from dual union all
select '10258932' a, '7'  b, '784,88'  c, '23,19'  d,  'E'    e from dual union all
select '10258932' a, '8'  b, '100,88'  c, '11,11'  d,  'E'    e from dual union all
select '10258932' a, '9'  b, '300,88'  c, '22,22'  d,  'F'    e from dual union all
select '10258932' a, '10' b, '468,25'  c, '78,33'  d,  'F'    e from dual
)
--
-- end-of-test-data
--
select a, b, c, d, e,
       num_distinct_values
from
(
  select a, b, c, d, e,
         count(distinct e) over (partition by a) num_distinct_values
  from my_table)
where num_distinct_values <= 1
order by 1
/

The county notes how the distinct non-null values in column e for each one.
Where clause then selects those with count = 0 (this is where e = null) and count = 1.
If the number is greater than 1, different values exist for this id.

Choose what fits your condition ;-)

break the values of columns in separate lines

Hello

I have a query that returns the values of 4 columns of it. Of this, a column can have many values it contains separated by semocolon (-). The requirement now is if this particular column has more than one value, separated by the semocolon (-), then the query must return each of them separately as a new line and keep the rest, column values as common for each of these values of this colummn.

For example, select...

output:
col1 col2 col3
Risk of BB IE6 71 of the PACK_PWRMART file. DELETE_POSITION_WO_RISK ($FEEDINSTANCEID); PACK_PWRMART. DELETE_INSTRU_REL_POS_WO_RISK; QC 8401

COL4
2010-02-11 09:29:45 December 28, 2010

Now, col3 has 3 values separated by (;). Therefore, the output of the query

col1          col2                           col3( now split with 3 values)                                                          col4
71     BB IE6 risk file     PACK_PWRMART.DELETE_POSITION_WO_RISK($FEEDINSTANCEID)        2/11/2010 9:29:45 AM 28-Dec-2010     
72     BB IE6 risk file     PACK_PWRMART.DELETE_INSTRU_REL_POS_WO_RISK            2/11/2010 9:29:45 AM 28-Dec-2010     
73             BB IE6 risk file             QC 8401                                                                                       2/11/2010 9:29:45 AM 28-Dec-2010

The solution that I thought:

I don't know if this can be achieved in a single SQL statement, so I thought I should write a block of code that will use the collection separate each of the value of this column and use some sort of record of object to insert the entire line in it 3 times (or the number of times as much as the values in this column) , but without success.

Something like this:

declare
type feed_name is table of varchar2(4000);--feed_static.feed_description% type INDEX BY PLS_INTEGER;
feed_nm feed_name;
param_val feed_name;
dat feed_name;
feed_desc feed_name;
str number;
str1 varchar2(1000);
begin
select feed_description as "Feed Name" ,
       param_value as "Post_Proc Prarameter",
       '28-Dec-2010' as "COB date",
       ft.feed_type_description as "Feed Type"    
BULK COLLECT INTO feed_nm,param_val,dat,feed_desc
from feed_static fs, feed_parameter fp, feed_type ft 
where fs.feed_id =fp.feed_id
and fs.feed_type_id = ft.feed_type_id
and fs.feed_type_id not in (4,15,16,25)
and fp.param_name  = 'postproc'
and fp.close_action_id is null
and fs.feed_description in ('BB 0J0 risk file')--,'BB 1J1 risk file')
order by  fs.feed_type_id,feed_description;
FOR i IN 1..param_val.count
loop
    select length((param_val(i))) - length(replace(param_val(i),';')) into str from dual;
           dbms_output.put_line(str); 
           for j in 1..str
               loop
                    SELECT SUBSTR(param_val(i), 1 ,INSTR(param_val(i), ';', 1, j)-1) into str1 FROM dual;
                    dbms_output.put_line(str1); 
                    str1:=' ';
               end loop;
end loop;    
end;

But the "dbms_output.put_line (str1); does not give me o/p as required except the first value (assume thr are the 3 values of the column).
Also I'm not sure of what I'm trying to do is in the right direction or not.

I have not used any RECORDING still to poulate values separated again (I'm not sure about hoe to use)

Cam anyone help me please with this code.

Rgds,
Aashish

Try this:

select feed_description as "Feed Name" ,
       param_value as "Post_Proc Prarameter",
       '28-Dec-2010' as "COB date",
       --ft.feed_type_description as "Feed Type" ,
       regexp_substr( ft.feed_type_description,'[^;]+',1,rown.seq)
from feed_static fs, feed_parameter fp, feed_type ft, (SELECT LEVEL seq FROM DUAL
                    CONNECT BY LEVEL<=max(length(ft.feed_type_description)- length(replace(ft.feed_type_description,';'))))rown
where fs.feed_id =fp.feed_id
and fs.feed_type_id = ft.feed_type_id
and fs.feed_type_id not in (4,15,16,25)
and fp.param_name  = 'postproc'
AND rown.seq<=(length(ft.feed_type_description)- length(replace(ft.feed_type_description,';')))
and fp.close_action_id is null
and fs.feed_description in ('BB 0J0 risk file')--,'BB 1J1 risk file')
order by  fs.feed_type_id,feed_description;

regexp_count started in 11g. It counts the number of models in a particular string. I replaced by length - length (repalce).
You can replace the word 'max (length (ft.feed_type_description) - length (replace(ft.feed_type_description,';'))))' with 10 or 15, i.e. the maximum number of semicolons in the chain

How to remove duplicates in two separate lines in SQL?

Hello

I need to identify the data in double with two separate lines. See my below data structure.
INDEX OF NAMES NAME_1 START_DATE END_DATE FLAG VALUE
B ON 275 SE 13/12/2005-31/12/2010 1
SELF 1 375
SELF 1 475
ON 275 SE 2 B 13/12/2005-31/12/2010
SELF 2 375
SELF 2 475
ON SE 175 13/12/2006 AND 31/12/2010 B 3
SELF 3 375
SE ON 3 475

This is my sample data. Here, the data is mirrored with different index columns. INDEX 1 and 2 contains the same combination group. I so need to identify one of the combination double (i.e. INDEX 1 or 2). Anyone can come up with the exact solution?

Thank you

I thought it had to do with the order of the INDEX field. The query checks only the value of the previous INDEX value.

I think I found what the problem was.

Try this:

with test_table as
(select 'SUR' NAME, 'SE' NAME_1, '275' VALUE, to_date('13/12/2005','DD/MM/YYYY') START_DATE, to_date('31/12/2010','DD/MM/YYYY') END_DATE, 'B' FLAG, 1 IND from dual union all
select 'SUR', 'SE', '375', null, null, 'A', 1 from dual union all
select 'SUR', 'SE', '475', null, null, 'A', 1 from dual union all
select 'SUR', 'SE', '275', to_date('13/12/2005','DD/MM/YYYY') ,to_date('31/12/2010','DD/MM/YYYY') ,'B', 2 from dual union all
select 'SUR', 'SE', '375', null, null, 'A', 2 from dual union all
select 'SUR', 'SE', '475', null, null, 'A', 2 from dual union all
select 'SUR', 'SE', '175', to_date('13/12/2006','DD/MM/YYYY'), to_date('31/12/2010','DD/MM/YYYY'), 'B', 3 from dual union all
select 'SUR', 'SE', '375', null, null, 'A', 3 from dual union all
select 'SUR', 'SE', '475', null, null, 'A', 3 from dual union all
select 'SUR', 'SE', 'IPT', to_date('13/12/2005','DD/MM/YYYY'), to_date('31/12/2010','DD/MM/YYYY'), 'B', 100 from dual union all
select 'SUR', 'SE', '375',null, null, 'A', 100 from dual union all
select 'SUR', 'SE', 'AO', null, null, 'A', 100 from dual union all
select 'SUR', 'SE', 'IPT', to_date('13/12/2005','DD/MM/YYYY'), to_date('31/12/2010','DD/MM/YYYY'), 'B', 218 from dual union all
select 'SUR', 'SE', '375' ,null, null,'A' ,218 from dual union all
select 'SUR', 'SE', 'AM' ,null, null,'A' ,218 from dual union all
select 'SUR', 'SE', 'IPT' , to_date('13/12/2005','DD/MM/YYYY'), to_date('31/12/2008','DD/MM/YYYY'), 'B', 334 from dual union all
select 'SUR', 'SE', '375' ,null, null,'A', 334 from dual union all
select 'SUR', 'SE', 'AP',null, null, 'A', 334 from dual union all
select 'SUR', 'SE', 'IPT' , to_date('01/01/2009','DD/MM/YYYY'), to_date('31/12/2010','DD/MM/YYYY'), 'B', 335 from dual union all
select 'SUR', 'SE', '375' ,null, null,'A', 335 from dual union all
select 'SUR', 'SE', 'AP',null, null, 'A', 335 from dual union all
select 'SUR', 'SE', 'IPT' , to_date('13/12/2005','DD/MM/YYYY'), to_date('31/12/2010','DD/MM/YYYY'), 'B', 336 from dual union all
select 'SUR', 'SE', '375' ,null, null,'A', 336 from dual union all
select 'SUR', 'SE', 'AP',null, null, 'A', 336 from dual)
SELECT name,
       name_1,
       value,
       start_date,
       end_date,
       flag,
       ind,
       MIN(diff_flag) over(PARTITION BY NAME, NAME_1, IND) min_flag
  FROM (SELECT t3.*,
               CASE
                 WHEN lag(VALUE, lag_flag)OVER(ORDER BY IND, start_date, VALUE) != VALUE AND flag = 'A' THEN
                  1
                 WHEN (lag(value, lag_flag)OVER(ORDER BY IND, start_date, VALUE) != value
                    OR lag(start_date, lag_flag) OVER(ORDER BY IND, start_date, VALUE) != start_date
                    OR lag(end_date, lag_flag) OVER(ORDER BY IND, start_date, VALUE) != end_date)
                    AND flag = 'B' THEN
                  1
                 WHEN lag_flag = 0 THEN
                  1
                 ELSE
                  2
                 END diff_flag
          FROM (SELECT t2.*,
                       first_value(lag_count_flag) OVER(PARTITION BY NAME, NAME_1, IND
                                                            ORDER BY NAME, NAME_1, start_date) lag_flag
                  FROM (SELECT t1.*,
                               CASE
                                 WHEN FLAG = 'B' THEN
                                  lag(COUNT, 1, 0)
                                  OVER(ORDER BY IND, start_date, VALUE)
                               END lag_count_flag
                          FROM (SELECT t.*,
                                       COUNT(*) over(PARTITION BY NAME, NAME_1, IND) COUNT,
                                       row_number() over(PARTITION BY NAME, NAME_1, VALUE,
                                                                      START_DATE, END_DATE, FLAG
                                                             ORDER BY IND) row_number
                                  FROM test_table t) t1) t2) t3) t4
 ORDER BY ind,
          start_date,
          VALUE;

Published by: Manuel Vidigal on Abr 13, 2009 18:18

Retrun separate lines with duplicate values

Similar Questions

Maybe you are looking for