«

Hi,

I hope you may help me.

Please check the attached text file.
Actually its a report file with some headers information
and them report is in tabular format. I want to parse each row
and get the values.

I think strtok won't work here.

There are no tabs in between the column values they are spaces.
Column values them selves contain spaces but they are single spaces.
Minimum two spaces are there between column values.

can you suggest some algorithm in C or any other method to get each
column value.

In excel I try to create a delimited file but its unsuccessful.

Thanks,

lee


following content is in a text file. Copy in a text file
Keep wordwrap off
-------------------------------------------------------------------------------




DATE: 04/07/2007

TMK

MTC_PROC REPORT

Area: KTO
WRKPCK: KTO


PROCESS PNAME FEANAME CEB MCO DPNUM NM_ITM
ERROR
TXT_GUI ERROID
_______ ___________ _____________ ___ ___ ______ __________
________________________________________________
__________________________________________________ __________ ______

RECKKON 206 DEVICE 007 997 532533 532533
Invalid source. BREAK needed. Count: 1501-1505
Source: -98 0201

RECKKON U206 DEVICE 007 997 532533 532533
Invalid source. BREAK needed. Count:
1421,1726-1730 Source: -98 0201

RECKKON F77 CROSS CNNCT 009 997 520624 520624
Feeder feature. Count: AM992,651-900
Source: 520619 7310

RECKKON F1727 X-MIC 009 997 521184 521184
Sourced Out. Count: SA5211,1-1800
Source: 0 0206

RECKKON F115 CROSS CNNCT 009 997 522306 522306
provide feed. Count:
1400,1001-1100 Source: 522333 7310

On 11 Apr, 11:44, yang__...@ausi.com wrote:
> Hi,
>
> I hope you may help me.
>
> Please check the attached text file.
> Actually its a report file with some headers information
> and them report is in tabular format. I want to parse each row
> and get the values.
>
> I think strtok won't work here.
>
> There are no tabs in between the column values they are spaces.
> Column values them selves contain spaces but they are single spaces.
> Minimum two spaces are there between column values.
>
> can you suggest some algorithm in C or any other method to get each
> column value.

You know what data is in what character positions, so what's wrong
with extracting the data on that basis?

strncpy() would do for starters. More complex code could wrap this to
strip trailing spaces, convert to numeric format, etc...

wrote:
> Hi,
>
> I hope you may help me.
>
> Please check the attached text file.

No, thank you. Do not post attachments to text newsgroups.

In article <>,
Martin Ambuhl <> wrote:

>> Please check the attached text file.

>No, thank you. Do not post attachments to text newsgroups.

In fact he didn't post an attachment. He just appended the text to
his message.

-- Richard
--
"Consideration shall be given to the need for as many as 32 characters
in some alphabets" - X3.4, 1963.

Richard Tobin wrote:
> In article <>,
> Martin Ambuhl <> wrote:
>
>>> Please check the attached text file.
>
>> No, thank you. Do not post attachments to text newsgroups.
>
> In fact he didn't post an attachment. He just appended the text to
> his message.

As silly as it may be, I tend to believe posters unless this has already
been shown to be an error. At the point that (an
attribution you unaccountably removed) wrote that there was an attached
file, there was no reason to doubt his word. Silly me.

In the same way, if I encountered your post without known earlier
context, I might think that I had written
>>> Please check the attached text file.
I would wonder when I did it, and why there were extra '>'s, but you
claimed that I wrote, so I must have.

You have two formats, one for the header and one for the body.

Write a function to parse the header and another function to parse the
body.

Do not write one function to do both jobs. It is also easy to
recognize the transition between header and body so write a
controlling function that sees the header and calls the header parser
and then sees the body and calls the body parser.

Your question is really more appropriate for news:comp.programming
since you do not have a C question but a programming one.

Personally, I would parse the header and then throw the body into a
text file and use an ODBC text file connection. But that's neither
here nor there and it's not even topical on news:comp.lang.c

bcnu

In article <>,
Martin Ambuhl <> wrote:

>>>> Please check the attached text file.

>>> No, thank you. Do not post attachments to text newsgroups.

>> In fact he didn't post an attachment. He just appended the text to
>> his message.

>As silly as it may be, I tend to believe posters unless this has already
>been shown to be an error. At the point that (an
>attribution you unaccountably removed) wrote that there was an attached
>file, there was no reason to doubt his word. Silly me.

"Attached" doesn't only mean "MIME attachment". The phrase "the
attached X" has been around for a long time.

>In the same way, if I encountered your post without known earlier
>context, I might think that I had written
> >>> Please check the attached text file.
>I would wonder when I did it, and why there were extra '>'s, but you
>claimed that I wrote, so I must have.

When you become familiar with Usenet quoting conventions, you will not
have this problem.

-- Richard
--
"Consideration shall be given to the need for as many as 32 characters
in some alphabets" - X3.4, 1963.

Richard Tobin wrote:
> In article <>,
> Martin Ambuhl <> wrote:

>> In the same way, if I encountered your post without known earlier
>> context, I might think that I had written
>>>>> Please check the attached text file.
>> I would wonder when I did it, and why there were extra '>'s, but you
>> claimed that I wrote, so I must have.
>
> When you become familiar with Usenet quoting conventions, you will not
> have this problem.

When you become familiar with Usenet quoting conventions, you will no
longer snip attributions aways from text your are quoting. Your
practice is at least cavalier and sloppy; it is possibly dishonest as well.

Richard Tobin wrote:
>
> When you become familiar with Usenet quoting conventions, you will not
> have this problem.
>
Your conventions, or everyone else's?

--
Ian Collins.

user923005 wrote:

> You have two formats, one for the header and one for the body.

I do?

> Write a function to parse the header and another function to parse the
> body.

Why would I want to do that?

> Do not write one function to do both jobs. It is also easy to
> recognize the transition between header and body so write a
> controlling function that sees the header and calls the header parser
> and then sees the body and calls the body parser.

If you say so!

> Your question is really more appropriate for news:comp.programming
> since you do not have a C question but a programming one.

I didn't know I had a question.

> Personally, I would parse the header and then throw the body into a
> text file and use an ODBC text file connection. But that's neither
> here nor there and it's not even topical on news:comp.lang.c

You're the one who brought it up.


Unless, perhaps, this was meant as a reply to some OTHER person. As
there's no quotes are attributions, that's rank speculation on my part.




Brian