Shell Programming and Scripting

OKAY----

Here's what I must do.
I have two files. I need to compare the two files such as with the diff command. I am adding FILENEW to FILEOLD

If fields $1, $2, $5, and 6 are the same, then I don't want to add FILENEW records to FILEOLD.
If they are not, then append the lines.

Is this possible in the UNIX scripting world? It seems that this could be done in AIX. I also have Perl as a last resort. Looking hard, wondered if anyone has done something like this before.

it may be sickboys entry is similar...on phone will check when I get off

Does anybody want to take a crack at this?
My slow brain is simply not understanding.

If for example:

awk 'BEGIN{FS=OFS=","}
NR==FNR{a[$1]=$0; next}
a[$1]{$0=$0 RS a[$1]}{print}' file2 file1

Adds a line to file1 if the first field matches in both files.
How do you only print the line if say, fields1, 2, 5, and 6 match both files?
And of course, I would have to reverse that (i.e., only add the lines to file1 if fields 1, 2, 5, and 6 are NOT matched between file1 and file2)

Obviously the first FNR line is file2, a[$1] being the first field.
The second line then matches a[$1] and appends a record to the output of file1.

Thanks much for any assistance.

File1:
----------------------------
apples,pears,bananas,oranges,grapefruit,watermelon,pineapple
epples,pears,bananas,oranges,grapefruit,watermelon,pineapple
epples,paars,bananas,oranges,grapefruit,watermelon,pineapple
graby,epples,bananas,oranges,grapefruit,watermelon,pineapple


File2:
----------------------------
pearsrs,apples,bananers,orangas,grapefruit,graby,pineapples
apples,pears,bandanas,orangatangs,grapefruit,watermelon,wineapple



Wanted output: (File1: appended: only add records when fields 1, 2, 5 and 6 DO NOT match)
-----------------------------
apples,pears,bananas,oranges,grapefruit,watermelon,pineapple
epples,pears,bananas,oranges,grapefruit,watermelon,pineapple
epples,paars,bananas,oranges,grapefruit,watermelon,pineapple
graby,epples,bananas,oranges,grapefruit,watermelon,pineapple
pearsrs,apples,bananers,orangas,grapefruit,graby,pineapples

nawk -f jeff.awk FILENEW FILEOLD

jeff.awk:

Code:

{ idx = $1 SUBSEP $2 SUBSEP $5 SUBSEP $6 }
NR==FNR { fnew[ idx ] = $0; next }
{
   if ( !(idx in fnew) )
      $0 = $0 ORS fold[idx]
}
1

Basically what this is is a logfile that has alerts from a program.
The entries get dropped into a file continually. But some entries have the same date, transaction ID, location, alert code and keep getting dropped in over and over (like until the problem is fixed)

The awk filter is to create a 2nd file that just adds new messages if they are not the same date, ID, location, code, and run on a continual basis filtering these out.

Thanks
But something is awry.

OKAY, If I do:
nawk -f jeffawk file2 file1
------------------------
apples,pears,bananas,oranges,grapefruit,watermelon,pineapple
epples,pears,bananas,oranges,grapefruit,watermelon,pineapple

epples,paars,bananas,oranges,grapefruit,watermelon,pineapple

graby,epples,bananas,oranges,grapefruit,watermelon,pineapple

This is basically a cat of file1.


If I do:
nawk -f jeffawk file1 file2
I get the following:
-----------------------
pearsrs,apples,bananers,orangas,grapefruit,graby,pineapples

apples,pears,bananas,oranges,grapefruit,watermelon,pineapple



The first line, 'pearsrs', is right on the money. But this is basically a cat of file2.


Not sure
? ? ? ?