Return to Question

... Edited to final on 08/30/2014 7:00 PM

Replacing cut with sed seems to work. Firebird field input still needs to be escaped for semi-colons (') with another of the same added, and the current comma IFS delimiter in data files still has to be replaced with non-ASCII. Instead of recursive file lists this script is still a single-file input. echo should probably be replaced by printf, and a whole lot of more...

Here is the final script excluding the shebang:

renice -n 19 $$ >> Radix_en_log.txt;

idx=32768; dbt='Radix_en';
cat Domains_en.txt; cat Tables_en.txt;

while read;     do
gid="$idx,'$(uuidgen)'";

IFS=,; set -f;
echo "insert into $dbt values ($gid";
    for field in $REPLY
    do  
        printf '%s' ",$field,'"; printf '%s' "$field" | md5sum;
    done | sed "s/[ ][ ][-]/\'/g"; printf '%s\n' ");";
    idx=$((idx+1));
        done < Data.txt;
echo "commit;";

Here is the output:

create domain ...;
create domain ...;
commit;

create table ( ... );
create table ( ... );
commit;

insert into Radix_en values (32768,'2f675b86-b2b4-4e52-b000-e6a8cf0f3dca'
,'NUMBER','722d79c16b51fe86610972b8d950038c'
,'US_EN','b63fb39e32b062c76694bec58c4f8c67'
,'LATIN','fd6f27a3c59111fc2a0b5e452595ef3d'
,'GREEK','c081310697bb6b7d7bed5034824e2530'
,'GERMAN','15db1d0e1b0861d8ac1f391db801493a'
);
insert into Radix_en values (32769,'e2afcd65-9a1b-49e3-baf1-74b0619a4776'
,0,'cfcd208495d565ef66e7dff9f98764da'
,'zero','01b40535afbfd9611e910f58f4ab5146'
,'nulla','584edd0b6638798dee53e2c23e84e2d1'
,'μηδέν','cd3ed2f1039ed8668b4d48e742bd2e5b'
,'Null','e0a93a9e6b0eb1688837d8bab9b4badb'
);
insert into Radix_en values (32770,'f51b72eb-d64f-4e9e-ab49-8954df9505cd'
,1,'c4ca4238a0b923820dcc509a6f75849b'
,'one','7e31533231a12e4a560a18ac8cd02112'
,'Unum','05d92bcbffbf59b375f25945e9af2dd0'
,'ένα','826f5e2d5ba7ace48f4d6fe3c5e2925f'
,'ein','dcc09a2cb665ca332d1689cb11aff592'
);
commit;

Thanks!

Sandor

renice -n 19 $$ > /dev/null; #for now

idx=32768;
dbt='Radix_en';

cat Domains_en.txt;
cat Tables_en.txt;

while read; 
do

gid="$idx,'$(uuidgen)'";

IFS=,;
echo "insert into $dbt values ($gid";
    for field in $REPLY
    do  
    printf '%s' ",$field,'";
    printf '%s' "$field" | md5sum;
    done | cut -d\  -f1;
    

echo "$var);";

idx=$((idx+1));

done < Data.txt;

renice -n 19 $$ > /dev/null; #for now

idx=32768;
dbt='Radix_en';

cat Domains_en.txt;
cat Tables_en.txt;

while read; 
do

gid="$idx,'$(uuidgen)'";

IFS=,; set -f # no gobbling!
echo "insert into $dbt values ($gid";
    for field in $REPLY
    do  
    printf '%s' ",$field,'";
    printf '%s' "$field" | md5sum;
    done | cut -d\  -f1;
    

echo "$var);";

idx=$((idx+1));

done < Data.txt;

... Edited to add more on 08/23/2014 3:00 AM

Thanks to mikeserv using IFS is working, so my scripts now looks like this:

renice -n 19 $$ > /dev/null; #for now

idx=32768;
dbt='Radix_en';

cat Domains_en.txt;
cat Tables_en.txt;

while read; 
do

gid="$idx,'$(uuidgen)'";

IFS=,;
echo "insert into $dbt values ($gid";
    for field in $REPLY
    do  
    printf '%s' ",$field,'";
    printf '%s' "$field" | md5sum;
    done | cut -d\  -f1;
    

echo "$var);";

idx=$((idx+1));

done < Data.txt;

The output is great, the line breaks are making text edit/search much easier while Firebird is still happy, except one thing..

Here is the output:

create domain ...;
create domain ...;
commit;

create table ( ... );
create table ( ... );
commit;

insert into Radix_en values (32768,'303f8957-57cf-4485-ace4-d21c7cf144e6'
,'NUMBER','722d79c16b51fe86610972b8d950038c
,'US_EN','b63fb39e32b062c76694bec58c4f8c67
,'LATIN','fd6f27a3c59111fc2a0b5e452595ef3d
,'GREEK','c081310697bb6b7d7bed5034824e2530
,'GERMAN','15db1d0e1b0861d8ac1f391db801493a
);
insert into Radix_en values (32769,'e7fdf095-d31c-4c59-a23b-7ea67db7aefb'
,0,'cfcd208495d565ef66e7dff9f98764da
,'zero','01b40535afbfd9611e910f58f4ab5146
,'nulla','584edd0b6638798dee53e2c23e84e2d1
,'μηδέν','cd3ed2f1039ed8668b4d48e742bd2e5b
,'Null','e0a93a9e6b0eb1688837d8bab9b4badb
);
insert into Radix_en values (32770,'a21916b5-2a05-4656-ad4e-c8cfee1abfcc'
,1,'c4ca4238a0b923820dcc509a6f75849b
,'one','7e31533231a12e4a560a18ac8cd02112
,'Unum','05d92bcbffbf59b375f25945e9af2dd0
,'ένα','826f5e2d5ba7ace48f4d6fe3c5e2925f
,'ein','dcc09a2cb665ca332d1689cb11aff592
);
commit;

The md5 hash is missing a delimiter at the end, and I have no idea how to negotiate the output with the pipes. What is it I am not understanding here?

Since in this particular case the data fields are going to hold code for programmable ICs no extra characters are going to be acceptable in the checksum between the delimiters, and so far it looks like so. Again, the code is in ASCII and my delimiter is going be something that is not part of ASCII so it is safe. Firebird also going to reject anything not ASCII.

If you would be so kind to point me to how to finish this script, as I am already banging my head against some new issues IFS just showed me (yes, file paths on Windows). I'll try and see how this script is going to work with pure ASCII, then I would like to move on and make another post about some more issues.

Thanks again for your help!

Sandor