Validating a string with sscanf

[Richard]

I'm validating a date and time string which must be EXACTLY of the format

yy-mm-dd hh:mm:ss

and extracting the six numeric values using sscanf.

I'm using the format string "%2u-%2u-%2u %2u:%2u:%2u" which works, but it
allows each numeric field to be either 1 or 2 digits in length. Also the man
page for sscanf says that preceeding white space before a numeric is
ignored.

But I need the input string to be EXACTLY of the above format (each numeric
is exactly 2 digits long and there is no white space) otherwise the input
string should be rejected.

What is the best way of validating the input string against an EXACT format
string and exctracting the six numeric values, or must I manually examine
every character of the input string?

[Sean G. McLaughlin]

Re: Validating a string with sscanf

Richard wrote:

I'm validating a date and time string which must be EXACTLY of the format
>
yy-mm-dd hh:mm:ss

[snip]

What is the best way of validating the input string against an EXACT
format string and exctracting the six numeric values, or must I manually
examine every character of the input string?

If you want to know the absolute best way, then you have one of three
courses of action:

1. Use a DSL that is suited to this task and that generates C code.
lex/flex is quite good.

2. Use a library that provides regexes, and validate with that. Regex
libraries are a dime a dozen, too. For example, they are in POSIX.2
(see 'man 3 regex' on Unix).

3. Drop C and use a language that has (1) or (2) built-in, like Perl or Awk
or what have you.


These strategies have disadvantages, though. (1) means you need to learn a
new language, (2) means you now have two problems ;-), and (3) might not be
an option depending on your situation. At this point you'll have to look
into finite state automata.

[Ben Bacarisse]

Re: Validating a string with sscanf

"Richard" <null@null.co.u kwrites:

I'm validating a date and time string which must be EXACTLY of the format
>
yy-mm-dd hh:mm:ss
>
and extracting the six numeric values using sscanf.
>
I'm using the format string "%2u-%2u-%2u %2u:%2u:%2u" which works, but
it allows each numeric field to be either 1 or 2 digits in
length. Also the man page for sscanf says that preceeding white space
before a numeric is ignored.
>
But I need the input string to be EXACTLY of the above format (each
numeric is exactly 2 digits long and there is no white space)
otherwise the input string should be rejected.
>
What is the best way of validating the input string against an EXACT
format string and exctracting the six numeric values, or must I
manually examine every character of the input string?

If all you want (at this stage) is syntactic correctness, then you
could test that sscanf returns 6, strlen(data) == 17 and that data
contains only one character for which isspace is true.

--
Ben.

[Richard]

Re: Validating a string with sscanf


"Ben Bacarisse" <ben.usenet@bsb .me.ukwrote in message
news:871w4r4dhi .fsf@bsb.me.uk. ..

"Richard" <null@null.co.u kwrites:
>

>I'm validating a date and time string which must be EXACTLY of the format
>>
>yy-mm-dd hh:mm:ss
>>
>and extracting the six numeric values using sscanf.
>>
>I'm using the format string "%2u-%2u-%2u %2u:%2u:%2u" which works, but
>it allows each numeric field to be either 1 or 2 digits in
>length. Also the man page for sscanf says that preceeding white space
>before a numeric is ignored.
>>
>But I need the input string to be EXACTLY of the above format (each
>numeric is exactly 2 digits long and there is no white space)
>otherwise the input string should be rejected.
>>
>What is the best way of validating the input string against an EXACT
>format string and exctracting the six numeric values, or must I
>manually examine every character of the input string?

>
If all you want (at this stage) is syntactic correctness, then you
could test that sscanf returns 6, strlen(data) == 17 and that data
contains only one character for which isspace is true.

Thanks. As this is a one-off, I've ended up checking all 17 characters
individually, with isdigit being used for each character position where a
numeric is required, and then calling sscanf to decode it.

[CBFalconer]

Re: Validating a string with sscanf

Richard wrote:

>
I'm validating a date and time string which must be EXACTLY of
the format
yy-mm-dd hh:mm:ss
and extracting the six numeric values using sscanf.

I think you should make great efforts to use ISO approved date
forms, which means that the year field will contain 4 digits. This
will avoid any further year 2000 flaps until 10000 AD, at which
point neither of us will probably care.

--
[mail]: Chuck F (cbfalconer at maineline dot net)
[page]: <http://cbfalconer.home .att.net>
Try the download section.


** Posted from http://www.teranews.com **

[Jack Klein]

Re: Validating a string with sscanf

On Sun, 27 Apr 2008 13:14:21 -0400, CBFalconer <cbfalconer@yah oo.com>
wrote in comp.lang.c:

Richard wrote:

I'm validating a date and time string which must be EXACTLY of
the format
yy-mm-dd hh:mm:ss
and extracting the six numeric values using sscanf.

>
I think you should make great efforts to use ISO approved date
forms, which means that the year field will contain 4 digits. This
will avoid any further year 2000 flaps until 10000 AD, at which
point neither of us will probably care.

Speak for yourself, I plan on being around that long.

--
Jack Klein
Home: http://JK-Technology.Com
FAQs for
comp.lang.c http://c-faq.com/
comp.lang.c++ http://www.parashift.com/c++-faq-lite/
alt.comp.lang.l earn.c-c++

http://www.club.cc.cmu.edu/~ajo/docs/FAQ-acllc.html

[Peter Nilsson]

Re: Validating a string with sscanf

Richard wrote:

I'm validating a date and time string which must be EXACTLY
of the format
>
yy-mm-dd hh:mm:ss
>
and extracting the six numeric values using sscanf.
>
I'm using the format string "%2u-%2u-%2u %2u:%2u:%2u"
which works, but it allows each numeric field to be either 1 or
2 digits in length. Also the man page for sscanf says that
preceeding white space before a numeric is ignored.
>
But I need the input string to be EXACTLY of the above
format (each numeric is exactly 2 digits long and there
is no white space) otherwise the input string should be
rejected.
>
What is the best way of validating the input string against
an EXACT format string and exctracting the six numeric
values, or must I manually examine every character of
the input string?

Pretty much, though you can do it with scanf...

#include <stdio.h>

#define DIGIT "%1[0123456789]"
#define SPACE "%*1[ ]"

#define FORMAT \
DIGIT DIGIT "-" DIGIT DIGIT "-" DIGIT DIGIT \
SPACE \
DIGIT DIGIT ":" DIGIT DIGIT ":" DIGIT DIGIT

#define DIGIT_PAIR_TO_I NT(X) \
( (X ## 1[0] - '0') * 10 + (X ## 2[0] - '0') )

int main(int argc, char **argv)
{
if (argc)
while (argv++, --argc)
{
char Y1[2], Y2[2];
char M1[2], M2[2];
char D1[2], D2[2];
char h1[2], h2[2];
char m1[2], m2[2];
char s1[2], s2[2];

int r = sscanf( *argv,
FORMAT,
Y1, Y2, M1, M2, D1, D2,
h1, h2, m1, m2, s1, s2 );

if (r == 12)
{
int Y = DIGIT_PAIR_TO_I NT(Y);
int M = DIGIT_PAIR_TO_I NT(M);
int D = DIGIT_PAIR_TO_I NT(D);
int h = DIGIT_PAIR_TO_I NT(h);
int m = DIGIT_PAIR_TO_I NT(m);
int s = DIGIT_PAIR_TO_I NT(s);

printf(" match:");
printf(" %02d-%02d-%02d", Y, M, D);
printf(" %02d:%02d:%02d" , h, m, s);
printf("\n");
}
else
printf("no match: %s\n", *argv);
}

return 0;
}

--
Peter