split file contents into files

[Harry]

I have extracted some scripts from an Oracle database using
the Toad query below. The script body is of datatype HUGHCLOB;
which means each script may be from 50 - 200+ lines.
N.B. Toad is a tool to query oracle database.


select
script_name
, script_type
, '>>>>>'
, script_body
from
script_table
order by
script_type, script_name;

The Toad query returned a single text file like this below.

What I want is an shell script, say, AWK, to parse
this all-in-one TOAD output file, to split each script
into its own script files, with names being script_name1.txt,
script_name2.txt, ... etc.

Any help appreciated.

--
script_name1,TypeA,>>>>>,
;comment line 1
;comment line 2
<blank line>
script 1 line 1
script 1 line 2
....
scrtp 1 last line
script_name2,TypeB,>>>>>,
/*
* comments for script 2
* ...
*/
script 2 line 1
script 2 line 2
...
--

TIA

[Ed Morton]

based on the above, it seems like something like this would work:

awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}'

Ed.

[Harry]

Ed Morton wrote...

Ed,

I got the folloing error.
I'm on cygwin with GNU awk 4.0.0.

$ awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}' <
TOAD_outfile.wri
awk: cmd. line:1: (FILENAME=- FNR=1) fatal: expression for `>' redirection has
null string value

Any clue?

[Harry]

Harry wrote...
>
>Ed Morton wrote...
[...]

>>based on the above, it seems like something like this would work:
>>
>>awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}'
>
>Ed,
>
>I got the folloing error.
>I'm on cygwin with GNU awk 4.0.0.
>
>$ awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}' <
>TOAD_outfile.wri
>awk: cmd. line:1: (FILENAME=- FNR=1) fatal: expression for `>' redirection
has
>null string value
>
>Any clue?

Oooppp... my bad, I forgot to trim my sql statement at the top
of my file, TOAD_outfile.wri.

After removing them ... your solution worked perfectly.

Thanks

[Harry]

On May 31, 8:10 pm, Ed Morton <mortons...@gmail.com> wrote:
[...]

>
> based on the above, it seems like something like this would work:
>
> awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}'
>

The following line does not belong to the script body.
How could I skip this line showing up in the generated script file(s)?
--
script_name1,TypeA,>>>>>,
--

Thanks

[Ed Morton]

awk -F, '/>>>>>/ {close(file); file=$1 ".txt"; next} {print> file}'

Ed

[Harry]

On May 31, 8:10 pm, Ed Morton <mortons...@gmail.com> wrote:
[...]
>

> based on the above, it seems like something like this would work:
>
> awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}'

The following line does not belong to the script body.

How could I exclude it from the generated script files?

[Harry]

On Jun 1, 8:28 am, Harry <harryooopot...@hotmail.com> wrote:
[...]

> > awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print > file}'
>
> The following line does not belong to the script body.
> How could I exclude it from the generated script files?
> --
> script_name1,TypeA,>>>>>,
> --

OK, I got it with this.

awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {if ($3 != ">>>>>")
print $0 >> file}'< TOAD_outfile.wri

[Ed Morton]

you might want to check the response I posted this morning for a better solution.

Ed.

[Harry]

Ed Morton wrote...

Yes, Agree.

I saw your reply after I sent mime.

Thanks a lot.

[Ed Morton]

You're welcome. If you're going to be using awk, though, I recommend you get the
book "Effective Awk Programming, 3rd Edition" by Arnold Robins. There's a couple
of things in the script you posted above ("if (..)" and "print >>") that are red
flags to fundamental misunderstandings on awk semantics.

Ed.

[Harry]

Ed Morton wrote...

[...]

>>> On 6/1/2012 3:24 PM, Harry wrote:
>>>> [...]
>>>>>> awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print> file}'

[...]

>You're welcome. If you're going to be using awk, though, I recommend you get
the
>book "Effective Awk Programming, 3rd Edition" by Arnold Robins. There's a
couple
>of things in the script you posted above ("if (..)" and "print >>") that are
red
>flags to fundamental misunderstandings on awk semantics.

Ed. I will certainly take your advice to get myself more into awk.

BTW, I tried out (print > file) and it worked and I guess it has to do
with awk being working on stream.

On the "if (..)" part, let me guess ... I should not use $3 due to
the fact that not every line has $3 ... isn't that what I did wrong?

Or you meant something else?

Thanks

[Ed Morton]

On 6/2/2012 6:26 PM, Harry wrote:
> Ed Morton wrote...
>
> [...]
>
>>>> On 6/1/2012 3:24 PM, Harry wrote:
>>>>> [...]
>>>>>>> awk -F, '/>>>>>/ {close(file); file=$1 ".txt"} {print> file}'
> [...]
>> You're welcome. If you're going to be using awk, though, I recommend you get
> the
>> book "Effective Awk Programming, 3rd Edition" by Arnold Robins. There's a
> couple
>> of things in the script you posted above ("if (..)" and "print>>") that are
> red
>> flags to fundamental misunderstandings on awk semantics.
>
> Ed. I will certainly take your advice to get myself more into awk.
>
> BTW, I tried out (print> file) and it worked and I guess it has to do
> with awk being working on stream.

The issue is that the semantics of > and >> are different in awk than in shell.
You should be using > in your script.

>
> On the "if (..)" part, let me guess ... I should not use $3 due to
> the fact that not every line has $3 ... isn't that what I did wrong?
>
> Or you meant something else?
>

Something else. Awk scripts are made up of:

<condition> { <action> }

statements. So you naturally should put your conditions in the condition part
rather than the action part. i.e. instead of:

{ if (<condition>) { <action> } }

you should simply write:

<condition> { <action> }

unless there's a good reason not to do that (which in this case there is not).

> Thanks
>

You're welcome.

Ed.