Linux RedHat, Inserting lines into file with a list from other file.

Try this:

sed -i '/LOAD DATA/ r list.lst' test.ctl

This will add the lines in list.lst into test.ctl after the LOAD DATA line.

awk '/APPEND/ {l=$0; while ((getline < "list.lst") > 0) {print "INFILE",$0}; print l; next} {print}' test.ctl

woolmilkporc, your suggestion is ok, with a modification:

ls -1 Transfer_data* > lista.lst
line=$(cat test.ctl | grep -n 'APPEND' | grep -o '^[0-9]*')
line=$((line -0))                                                                   #here I do a change
     for f in $(cat  lista.lst)
        do
           sed  -i ${line}" i\INFILE $f" test.ctl
          # sqlldr epe/epe@xa21 control=test.ctl log=sqlldr.log
        done

Select allOpen in new window

But when test.ctl is filled, the order is distinct that lista.lst:
see lista.ctl

Transfer_data.20141118194103
Transfer_data.20141118194132

Select allOpen in new window

and see the order inside test.ctl

INFILE Transfer_data.20141118194132
INFILE Transfer_data.20141118194103

Select allOpen in new window

How I can reverse the order?
Thanks

It's quite normal that the order gets reversed - we read the list file line by line and insert that respective line just above "APPEND", i.e we read top-down and insert bottom-up.

We can use the "tac" utiity (on Linux only) instead of "cat" to reverse the order of the lines as they're presented to "for".
Line #4 of the last version you posted should then read:

 for f in $(tac  lista.lst)

Another option: Reverse the order in lista.lst by instructing "ls" to display the newest file first: "ls -t1 ...".

Obviously you should use just one of the above suggestions at a time!

By the way, why would you want to subtract "0" from "line"? The variable $line should be numeric ab ovo, so no need for an arithmetic NOP.

And how about my "awk" solution? Here the file is inserted as is above "APPEND", so the order will be kept automatically. Is it because "awk" cannot change "in place"?

Finally, the "sqlldr" command should not appear inside the "for" loop, because test.ctl will not contain all "INFILE" lines from lista.lst before the last iteration has completed!

The 'sed' suggestion I made does not reverse order and is much shorter as well ;) So why don't you use that?

woolmilkporc,

I appreciate your attention and suggestions. Finally, use the option "ls -t1", and so far it works.

Also, I took the loop sentence "sqlldr ...." was my mistake.

Thanks again.

Gerwin Jansen, your suggestion not insert lines in test.ctl

Thanks anyway

>> Gerwin Jansen, your suggestion not insert lines in test.ctl
Sure it did, did you try my suggestion at all?

[gerwin@localhost ~]$ cat test.ctl
LOAD DATA
APPEND
INTO TABLE "INT_OPERATIONS"
FIELDS TERMINATED BY '!'
OPTIONALLY ENCLOSED BY '"'
TRAILING NULLCOLS
(
SCADA_ID CHAR,
SCADA_SISTEMA CHAR,
SCADA_OPERATION_CLASS CHAR,
SCADA_OPERATION_SUBCLASS CHAR,
SCADA_OPERATION_DATE ,
SCADA_DATE_OFFSET ,
SCADA_OPERATION_TIMESTAMP "TO_TIMESTAMP(RPAD(SUBSTR(:SCADA_OPERATION_TIMESTAMP,1,21),24,'0'),'dd/mm/yyyy hh24:mi:ss.ff)')",
CERTA_PROCESS_STATUS ,
CERTA_PROCESS_DATE "TO_TIMESTAMP(RPAD(SUBSTR(:CERTA_PROCESS_DATE,1,21),24,'0'),'dd/mm/yyyy hh24:mi:ss.ff)')",
CERTA_OPERATION_ID ,
CERTA_NOTES CHAR,
INT_ID
 )
[gerwin@localhost ~]$ cat list.lst
Transfer_data.20141118194103
Transfer_data.20141118194132
[gerwin@localhost ~]$ sed -i '/LOAD DATA/ r list.lst' test.ctl
[gerwin@localhost ~]$ cat test.ctl
LOAD DATA
Transfer_data.20141118194103
Transfer_data.20141118194132
APPEND
INTO TABLE "INT_OPERATIONS"
FIELDS TERMINATED BY '!'
OPTIONALLY ENCLOSED BY '"'
TRAILING NULLCOLS
(
SCADA_ID CHAR,
SCADA_SISTEMA CHAR,
SCADA_OPERATION_CLASS CHAR,
SCADA_OPERATION_SUBCLASS CHAR,
SCADA_OPERATION_DATE ,
SCADA_DATE_OFFSET ,
SCADA_OPERATION_TIMESTAMP "TO_TIMESTAMP(RPAD(SUBSTR(:SCADA_OPERATION_TIMESTAMP,1,21),24,'0'),'dd/mm/yyyy hh24:mi:ss.ff)')",
CERTA_PROCESS_STATUS ,
CERTA_PROCESS_DATE "TO_TIMESTAMP(RPAD(SUBSTR(:CERTA_PROCESS_DATE,1,21),24,'0'),'dd/mm/yyyy hh24:mi:ss.ff)')",
CERTA_OPERATION_ID ,
CERTA_NOTES CHAR,
INT_ID
 )
[gerwin@localhost ~]$

Select allOpen in new window