Report program

[Former Member]

Hi everyone,

Consider the two fields person and ssn.For a particular ssn value if exists two or more different person values,then i want to put that details as a duplicate records to the internal table.Kindly help me in making this.Thanx in advance.

[Former Member]

Hi Satheesh,

you can assign the field-symbol to any data area. It does not matter if you declared your internal table with or without header line. I think you mean header line when you say work area. A work area usually is declared separately:

data:
  itab1 type table of whatever,
  itab_wa type whatever,
  itab2 type table of whatever with header line.
field-symbols:
  <whatever> type whatever.
* internal table without header line using work area
loop at itab1 into itab_wa.
  itab_wa-field1 = 'XYZ'.
  modify itab1 from itab_wa.
endloop.

* internal table with header line (implicit work area)
loop at itab2.
  itab2-field1 = 'XYZ'.
  modify itab2.
endloop.

* internal table without header line using field-symbol
loop at itab1[] assigning <whatever>.
  <whatever>-field1 = 'XYZ'.
endloop.

* internal table with header line using field-symbol
loop at itab2 assigning <whatever>.
  <whatever>-field1 = 'XYZ'.
endloop.

This is all the same. Internal tables with header line are not allowed in object-oriented context, i.e. declared within classes because the internal table has the same name as the header line. in non-OO-context you can distinguish between table body and headerline with the addition of square brackets [] for table body.

The field-symbols alternative is a lot faster. Using explicit (work area) or implicit header line means to copy all fields from body to header line/work area for each loop and copy all fields from header line/work area back to the boy using the modify statement.

Using the field symbols with LOOP...ASSIGNING means to have direct access to the actual record in table body.

Regards,

Clemens

[amit_khare]

Hi,

select * from PA0002 into table itab where perid = <var>

begda <= sy-datum

endda >= sy-datum.

endselect.

this will call all values in table itab now you can get your pernr there.

Regards,

Amit

[Former Member]

Hi Satheesh,

Plz use the following code.

Lets assume you have the data in LT_DATA.

Sort WA_DATA by SSN.

loop at LT_DATA into WA_DATA.
if lv_ssn equals to WA_DATA-SSN.
 lv_flag  = 1.
endif.
lv_ssn = WA_DATA-ssn.
At new ssn.
 if lv_flag = 1.
* --> Please place your duplicate related code here
 endif.
endat.
clear lv_flag.
endloop.

Regards

Bhupal Reddy

[Former Member]

Hi everyone,

Thanx for ur comments. one small correction there will be more ssn values in the table and for each entry i have to validate if there are more persons assigned an d there by i have to send the duplicate records for the single SSN value .In that case i can't assign the local variable as you all mentioned.Thanks.

[Clemenss]

Hi,

find duplicates by inserting into table with unique key. Something like that:

data:
* define unstructured hashed tables to check duplicates
  lt_ssn type hashed table of ssn
   with unique key table_line,
  lt_ssn_dup  type hashed table of ssn
   with unique key table_line.

loop at itab.
  insert itab-ssn into table  lt_ssn.
  if sy-subrc <> 0.
* insert fails because ssn already exists -> duplicate
      insert itab-ssn into table  lt_ssn_dup.
  endif.
endloop.
* now all duplicates are in lt_ssn_dup.
loop at itab.
  read table lt_ssn_dup transporting no fields
    with key table_line = itab-ssn.
  if sy-subrc = 0.
* process duplicate ssn here...
  else.
* process unique ssn here...
  endif.
endloop.

I would prefer to use LOOP AT itab ASSIGNING <field-symbol> but many colleagues are just frightened by field-symbols - you must get used to the performance gain...

Instead of TYPE ssn use the data element linked to field ssn.

Regards,

Clemens

[Former Member]

Hi Li,

Thanx for valuable information.Can i able to do the same with field symbols..I am new to this..just gimme the same example in terms of the field symbols please.Thanks in advance.

[Clemenss]

Hi Satheesh,

just a slight difference. Looping tables usually means a copy of the table record to the header line for each loop, using field-symbols just puts an address pointer to the actual record. Then you can access it directly. If (not here!) you modify a table, changing the field-symbols content means changing the table record - so no modify is will be used.

Now the code:

field-symbols:
  <itab> like itab.
loop at itab assigning <itab>.
  insert <itab>-ssn into table  lt_ssn.
  if sy-subrc <> 0.
* insert fails because ssn already exists -> duplicate
      insert <itab>-ssn into table  lt_ssn_dup.
  endif.
endloop.
* now all duplicates are in lt_ssn_dup.
loop at itab assigning <itab>.
  read table lt_ssn_dup transporting no fields
    with key table_line = <itab>-ssn.
  if sy-subrc = 0.
* process duplicate ssn here...
  else.
* process unique ssn here...
  endif.
endloop.

Note: If you use field-symbol looping at hashed or sorted table, you can not change the key fields - in debugger they are greyed out.

Have fun with field-symbols!

Regards

Clemens

[Li is the family name, so please Clemens or Mr. Li)

[Former Member]

Hi Clemens,

Thank you for ur valuable information.sorry for the mistake.:).I am using the internal with work area is it possible to assign the field symbol to a work area for the same example you gave me...

[Clemenss]

Hi Satheesh,

you can assign the field-symbol to any data area. It does not matter if you declared your internal table with or without header line. I think you mean header line when you say work area. A work area usually is declared separately:

data:
  itab1 type table of whatever,
  itab_wa type whatever,
  itab2 type table of whatever with header line.
field-symbols:
  <whatever> type whatever.
* internal table without header line using work area
loop at itab1 into itab_wa.
  itab_wa-field1 = 'XYZ'.
  modify itab1 from itab_wa.
endloop.

* internal table with header line (implicit work area)
loop at itab2.
  itab2-field1 = 'XYZ'.
  modify itab2.
endloop.

* internal table without header line using field-symbol
loop at itab1[] assigning <whatever>.
  <whatever>-field1 = 'XYZ'.
endloop.

* internal table with header line using field-symbol
loop at itab2 assigning <whatever>.
  <whatever>-field1 = 'XYZ'.
endloop.

This is all the same. Internal tables with header line are not allowed in object-oriented context, i.e. declared within classes because the internal table has the same name as the header line. in non-OO-context you can distinguish between table body and headerline with the addition of square brackets [] for table body.

The field-symbols alternative is a lot faster. Using explicit (work area) or implicit header line means to copy all fields from body to header line/work area for each loop and copy all fields from header line/work area back to the boy using the modify statement.

Using the field symbols with LOOP...ASSIGNING means to have direct access to the actual record in table body.

Regards,

Clemens