Join Lines

Prev: vpath w/implicit pattern rule, & non-basename target (GNU make)
Next: grepping from last file of a listing......

From: Janis Papanagnou on 6 Jun 2010 10:16

---

shulamitm wrote:
> On 6 יוני, 12:50, superpollo <ute...(a)esempio.net> wrote:
>> shulamitm ha scritto:
>>
>>
>>
>>
>>
>>> Hello,
>>> I'm looking for solution to the following problem:
>>> I have a file contains 2 fileds, but in a few rows, the second field
>>> go down (without spaces) to the next line.
>>> How can I join the splitted lines?
>>> For exmple:
>>> I need to change the following :
>>> ERROR1 1
>>> ERROR2 1
>>> WARNING1 0
>>> WARNING2
>>> 0
>>> WARNING3
>>> 0
>>> ERROR3 1
>>> To:
>>> ERROR1 1
>>> ERROR2 1
>>> WARNING1 0
>>> WARNING2 0
>>> WARNING3 0
>>> ERROR3 1
>>> Thanks in advance!
>> if you dont mind the whitespace:
>>
>> ~/superpollo$ cat fields.sh
>> #!/usr/bin/env bash
>> while read A B
>> do
>> echo -n $A" "
>> if [ -z "$B" ]
>> then
>> read B
>> echo $B
>> else
>> echo $B
>> fi
>> done
>> ~/superpollo$ cat fields
>> ERROR1 1
>> ERROR2 1
>> WARNING1 0
>> WARNING2
>> 0
>> WARNING3
>> 0
>> ERROR3 1
>> ~/superpollo$ ./fields.sh < fields
>> ERROR1 1
>> ERROR2 1
>> WARNING1 0
>> WARNING2 0
>> WARNING3 0
>> ERROR3 1
>> ~/superpollo$
>>
>> bye
>>
>> --
>> Il valore 96 sarebbe,Volendo, un ipercubo in i.-הסתר טקסט מצוטט-
>>
>> -הראה טקסט מצוטט-
>
> thanks!
> how can I do the same in korn shell or c shell?

While the proposed code is nothing I would suggest to use I also don't
see that the proposed code would not run in Kornshell; have you tried?

(And don't try to start using C shell for scripting.)

Janis

From: superpollo on 6 Jun 2010 10:58

---

Janis Papanagnou ha scritto:
> shulamitm wrote:
>> On 6 יוני, 12:50, superpollo <ute...(a)esempio.net> wrote:
>>> shulamitm ha scritto:
>>>
>>>
>>>
>>>
>>>
>>>> Hello,
>>>> I'm looking for solution to the following problem:
>>>> I have a file contains 2 fileds, but in a few rows, the second field
>>>> go down (without spaces) to the next line.
>>>> How can I join the splitted lines?
>>>> For exmple:
>>>> I need to change the following :
>>>> ERROR1 1
>>>> ERROR2 1
>>>> WARNING1 0
>>>> WARNING2
>>>> 0
>>>> WARNING3
>>>> 0
>>>> ERROR3 1
>>>> To:
>>>> ERROR1 1
>>>> ERROR2 1
>>>> WARNING1 0
>>>> WARNING2 0
>>>> WARNING3 0
>>>> ERROR3 1
>>>> Thanks in advance!
>>> if you dont mind the whitespace:
>>>
>>> ~/superpollo$ cat fields.sh
>>> #!/usr/bin/env bash
>>> while read A B
>>> do
>>> echo -n $A" "
>>> if [ -z "$B" ]
>>> then
>>> read B
>>> echo $B
>>> else
>>> echo $B
>>> fi
>>> done
>>> ~/superpollo$ cat fields
>>> ERROR1 1
>>> ERROR2 1
>>> WARNING1 0
>>> WARNING2
>>> 0
>>> WARNING3
>>> 0
>>> ERROR3 1
>>> ~/superpollo$ ./fields.sh < fields
>>> ERROR1 1
>>> ERROR2 1
>>> WARNING1 0
>>> WARNING2 0
>>> WARNING3 0
>>> ERROR3 1
>>> ~/superpollo$
>>>
>>> bye
>>>
>>> --
>>> Il valore 96 sarebbe,Volendo, un ipercubo in i.-הסתר טקסט מצוטט-
>>>
>>> -הראה טקסט מצוטט-
>> thanks!
>> how can I do the same in korn shell or c shell?
>
> While the proposed code is nothing I would suggest to use

can you explain why please?

bye

From: Janis Papanagnou on 6 Jun 2010 11:47

---

superpollo wrote:
> Janis Papanagnou ha scritto:
>> shulamitm wrote:
>>> On 6 יוני, 12:50, superpollo <ute...(a)esempio.net> wrote:
>>>> shulamitm ha scritto:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>> Hello,
>>>>> I'm looking for solution to the following problem:
>>>>> I have a file contains 2 fileds, but in a few rows, the second field
>>>>> go down (without spaces) to the next line.
>>>>> How can I join the splitted lines?
>>>>> For exmple:
>>>>> I need to change the following :
>>>>> ERROR1 1
>>>>> ERROR2 1
>>>>> WARNING1 0
>>>>> WARNING2
>>>>> 0
>>>>> WARNING3
>>>>> 0
>>>>> ERROR3 1
>>>>> To:
>>>>> ERROR1 1
>>>>> ERROR2 1
>>>>> WARNING1 0
>>>>> WARNING2 0
>>>>> WARNING3 0
>>>>> ERROR3 1
>>>>> Thanks in advance!
>>>> if you dont mind the whitespace:
>>>>
>>>> ~/superpollo$ cat fields.sh
>>>> #!/usr/bin/env bash
>>>> while read A B
>>>> do
>>>> echo -n $A" "
>>>> if [ -z "$B" ]
>>>> then
>>>> read B
>>>> echo $B
>>>> else
>>>> echo $B
>>>> fi
>>>> done
>>>> ~/superpollo$ cat fields
>>>> ERROR1 1
>>>> ERROR2 1
>>>> WARNING1 0
>>>> WARNING2
>>>> 0
>>>> WARNING3
>>>> 0
>>>> ERROR3 1
>>>> ~/superpollo$ ./fields.sh < fields
>>>> ERROR1 1
>>>> ERROR2 1
>>>> WARNING1 0
>>>> WARNING2 0
>>>> WARNING3 0
>>>> ERROR3 1
>>>> ~/superpollo$
>>>>
>>>> bye
>>>>
>>>> --
>>>> Il valore 96 sarebbe,Volendo, un ipercubo in i.-הסתר טקסט מצוטט-
>>>>
>>>> -הראה טקסט מצוטט-
>>> thanks!
>>> how can I do the same in korn shell or c shell?
>>
>> While the proposed code is nothing I would suggest to use
>
> can you explain why please?

Code that has the following properties that you use I consider suspect
and unreliable;

1. misleading shebang unnecessarily set to bash
2. using while/read loops
3. using read without setting IFS and option -r
4. using echo and option -n instead of printf
5. applying unquoted variables on many places
6. having lines of code duplicated unnecessarily
7. not supporting handling file arguments
(silently expecting data only on stdin)

Janis

>
> bye

From: superpollo on 7 Jun 2010 04:53

---

Janis Papanagnou ha scritto:
> superpollo wrote:
>> Janis Papanagnou ha scritto:
>>> shulamitm wrote:
>>>> On 6 יוני, 12:50, superpollo <ute...(a)esempio.net> wrote:
>>>>> shulamitm ha scritto:
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> Hello,
>>>>>> I'm looking for solution to the following problem:
>>>>>> I have a file contains 2 fileds, but in a few rows, the second field
>>>>>> go down (without spaces) to the next line.
>>>>>> How can I join the splitted lines?
>>>>>> For exmple:
>>>>>> I need to change the following :
>>>>>> ERROR1 1
>>>>>> ERROR2 1
>>>>>> WARNING1 0
>>>>>> WARNING2
>>>>>> 0
>>>>>> WARNING3
>>>>>> 0
>>>>>> ERROR3 1
>>>>>> To:
>>>>>> ERROR1 1
>>>>>> ERROR2 1
>>>>>> WARNING1 0
>>>>>> WARNING2 0
>>>>>> WARNING3 0
>>>>>> ERROR3 1
>>>>>> Thanks in advance!
>>>>> if you dont mind the whitespace:
>>>>>
>>>>> ~/superpollo$ cat fields.sh
>>>>> #!/usr/bin/env bash
>>>>> while read A B
>>>>> do
>>>>> echo -n $A" "
>>>>> if [ -z "$B" ]
>>>>> then
>>>>> read B
>>>>> echo $B
>>>>> else
>>>>> echo $B
>>>>> fi
>>>>> done
>>>>> ~/superpollo$ cat fields
>>>>> ERROR1 1
>>>>> ERROR2 1
>>>>> WARNING1 0
>>>>> WARNING2
>>>>> 0
>>>>> WARNING3
>>>>> 0
>>>>> ERROR3 1
>>>>> ~/superpollo$ ./fields.sh < fields
>>>>> ERROR1 1
>>>>> ERROR2 1
>>>>> WARNING1 0
>>>>> WARNING2 0
>>>>> WARNING3 0
>>>>> ERROR3 1
>>>>> ~/superpollo$
>>>>>
>>>>> bye
>>>>>
>>>>> --
>>>>> Il valore 96 sarebbe,Volendo, un ipercubo in i.-הסתר טקסט מצוטט-
>>>>>
>>>>> -הראה טקסט מצוטט-
>>>> thanks!
>>>> how can I do the same in korn shell or c shell?
>>> While the proposed code is nothing I would suggest to use
>> can you explain why please?
>
> Code that has the following properties that you use I consider suspect
> and unreliable;
>
> 1. misleading shebang unnecessarily set to bash

is #!/usr/bin/bash better? what should i put there instead?

> 2. using while/read loops

what's the alternative?

> 3. using read without setting IFS and option -r

is it necessary in this case? why should i do that?

> 4. using echo and option -n instead of printf

what's the problem with echo?

> 5. applying unquoted variables on many places

are unquoted vars dangerous?

> 6. having lines of code duplicated unnecessarily

like echo $B ?

> 7. not supporting handling file arguments
> (silently expecting data only on stdin)

hey! it was a quickie! ;-)

> Janis
>
>> bye

thanks!

bye

From: Janis Papanagnou on 7 Jun 2010 07:52

---

superpollo schrieb:
> Janis Papanagnou ha scritto:
>> superpollo wrote:
>>> Janis Papanagnou ha scritto:
>>>> shulamitm wrote:
>>>>> On 6 יוני, 12:50, superpollo <ute...(a)esempio.net> wrote:
>>>>>> shulamitm ha scritto:
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Hello,
>>>>>>> I'm looking for solution to the following problem:
>>>>>>> I have a file contains 2 fileds, but in a few rows, the second field
>>>>>>> go down (without spaces) to the next line.
>>>>>>> How can I join the splitted lines?
>>>>>>> For exmple:
>>>>>>> I need to change the following :
>>>>>>> ERROR1 1
>>>>>>> ERROR2 1
>>>>>>> WARNING1 0
>>>>>>> WARNING2
>>>>>>> 0
>>>>>>> WARNING3
>>>>>>> 0
>>>>>>> ERROR3 1
>>>>>>> To:
>>>>>>> ERROR1 1
>>>>>>> ERROR2 1
>>>>>>> WARNING1 0
>>>>>>> WARNING2 0
>>>>>>> WARNING3 0
>>>>>>> ERROR3 1
>>>>>>> Thanks in advance!
>>>>>> if you dont mind the whitespace:
>>>>>>
>>>>>> ~/superpollo$ cat fields.sh
>>>>>> #!/usr/bin/env bash
>>>>>> while read A B
>>>>>> do
>>>>>> echo -n $A" "
>>>>>> if [ -z "$B" ]
>>>>>> then
>>>>>> read B
>>>>>> echo $B
>>>>>> else
>>>>>> echo $B
>>>>>> fi
>>>>>> done
>>>>>> ~/superpollo$ cat fields
>>>>>> ERROR1 1
>>>>>> ERROR2 1
>>>>>> WARNING1 0
>>>>>> WARNING2
>>>>>> 0
>>>>>> WARNING3
>>>>>> 0
>>>>>> ERROR3 1
>>>>>> ~/superpollo$ ./fields.sh < fields
>>>>>> ERROR1 1
>>>>>> ERROR2 1
>>>>>> WARNING1 0
>>>>>> WARNING2 0
>>>>>> WARNING3 0
>>>>>> ERROR3 1
>>>>>> ~/superpollo$
>>>>>>
>>>>>> bye
>>>>>>
>>>>>> --
>>>>>> Il valore 96 sarebbe,Volendo, un ipercubo in i.-הסתר טקסט מצוטט-
>>>>>>
>>>>>> -הראה טקסט מצוטט-
>>>>> thanks!
>>>>> how can I do the same in korn shell or c shell?
>>>> While the proposed code is nothing I would suggest to use
>>> can you explain why please?
>>
>> Code that has the following properties that you use I consider suspect
>> and unreliable;
>>
>> 1. misleading shebang unnecessarily set to bash
>
> is #!/usr/bin/bash better? what should i put there instead?

If you're using just standard shell features without relying on any
bash'isms I'd suggest specify just that; standard sh.

Especially if the OP hasn't asked for a specific shell it's better
to not assume that he has a bash on his system, and to not give the
impression that bash would be required to execute your code.

>
>> 2. using while/read loops
>
> what's the alternative?

Any tool that does the loop implicitly. I wouldn't want to say here
that you should use something else, but code with while/read loops
are at least suspicious; in many cases it has been pointed out here
that this construct turned out to be inferior.

>
>> 3. using read without setting IFS and option -r
>
> is it necessary in this case? why should i do that?

The bourne shell family of shells show some often unwanted behaviour
WRT interpreting special characters and handling of field-splitting/
whitespace characters if used in default mode, and you need to take
some actions to handle it as needed. Those two details are typical;
this is regularily explained here, you may look that up if you missed
it.

>
>> 4. using echo and option -n instead of printf
>
> what's the problem with echo?

As well regularily explained here, and certainly also in the FAQ.

>
>> 5. applying unquoted variables on many places
>
> are unquoted vars dangerous?

Usually, yes. Similar to point 3; the default behaviour of unquoted
variable expansion is not well designed in the bourne shell family.
As a rule of thumb; always quote your variables, this is what you
really want in most cases.

>
>> 6. having lines of code duplicated unnecessarily
>
> like echo $B ?

Yes. It's trivial, I know, but I've added it since I was writing up
that list anyway, so I thought it couldn't hurt mentioning it.

>
>> 7. not supporting handling file arguments
>> (silently expecting data only on stdin)
>
> hey! it was a quickie! ;-)

Sure. But since you asked. :-)

The point is that a couple tools already do that implicitly. If you
considered above point 2. you might as well have it already handled
implicitly.

Janis

>
>> Janis
>>
>>> bye
>
> thanks!
>
> bye

First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: vpath w/implicit pattern rule, & non-basename target (GNU make)
Next: grepping from last file of a listing......