home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
HPAVC
/
HPAVC CD-ROM.iso
/
CLESSONS.ZIP
/
LESSON8
< prev
next >
Wrap
Text File
|
1986-02-07
|
25KB
|
478 lines
.NT
A NOTE ABOUT THE LESSONS in C
.b4-24R5C4
These were written while the author was ~Ilearning~N the language and since
.R6C4
they are ~Ifree~N ( to copy and/or distribute ) there is a money-back
.R7C4
guarantee on the accuracy of each and every statement in the lessons (!)
.R9C4
The ~Idisplay~N program was written ( in C ) in order to provide a vehicle
.R10C4
for displaying the lessons.
.R12C5
.B
P.J.Ponzo
.B
Dept. of Applied Math
.B
Univ. of Waterloo
.B
Ontario N2L 3G1
.K16,32
PonzoTUTOR
.WNT
FILES FILES FILES
.R4C1
Whereas ~b~Igetchar()~N and ~b~Iprintf()~N will get characters from the
keyboard and print to the screen (the 'standard' input/output), we can
ask other functions available in the C-library (and ones we write
ourselves) to communicate with a ~Ifile on disk~N. Consider the excerpt:
1 ~b~Imain(number,name) ~N
2 ~b~Iint number; ~N
3 ~b~Ichar *name[]; ~N
4 ~b~I{ ~N
5 ~b~I FILE *fp, *fopen(); ~N
6 ~b~I fp=fopen(name[1],"r"); ~N
7 ~b~I if (fp == NULL) { ~N
8 ~b~I printf("\nSorry, can't open %s",name[1]); ~N
9 ~b~I exit(0); ~N
10 ~b~I } ~N
11 ~b~I /* rest of program goes here */ ~N
.w
The above program is expected to read a file on disk. We compile/link
the program, giving it the name ~Iread~N. Then we type: ~Iread letter~N.
The arguments to ~b~Imain()~N are the ~b~Inumber~N ~I2~N and the ARRAY
of POINTERS ( ~b~Iname[]~N ) pointing to the strings: ~Iread~N and ~Iletter~N.
.WR10C1
~V2 int number; ~N
~V3 char *name[]; ~N
.R20C1
.R20C1
We declare the two arguments of ~b~Imain()~N to be ~b~Iint~N and ~b~Ichar *~N.
In our example, ~b~Inumber~N will have the value ~I2~N and ~b~Iname[0]~N will
be a pointer to the string ~Iread~N and ~b~Iname[1]~N will point to ~Iletter~N.
Now we are ready to begin ~b~I{~N our ~b~Imain()~N program ....
.WR10C1
2 ~b~Iint number; ~N
3 ~b~Ichar *name[]; ~N
4 ~b~I{ ~N
~V5 FILE *fp, *fopen(); ~N
.R20C1
.R20C1
To ~Iopen~N a disk file, we use ~b~IFILE~N which calls upon the C-library
to look after the mechanics of communicating with the operating system in
order to access the disk. After ~b~IFILE~N come two ~r~Ipointers~N (we
know they're pointers because of the ~b~I*~N ..right?).
.WR20C1
.R20C1
Of course, if we intend to use the C-library, we had better ~V#include~N it!
.WK10,60
stdio.h!
.R8C1
~V0 #include <stdio.h> ~N
.WR20C1
~b~Ifp~N is a ~r~Ipointer~N variable which points to a FILE.
~b~Ifopen()~N is a function which ~Ireturns~N a ~r~Ipointer~N to a FILE.
.WR13C1
5 ~b~I FILE *fp, *fopen(); ~N
~V6 fp=fopen(name[1],"r"); ~N
.R20C1
.R20C1
This calls upon the ~b~Ifopen()~N function to ~Iopen~N a file. The pointer
to the file ( which is ~Ireturn~Ned by ~b~Ifopen()~N ) is assigned to ~b~Ifp~N.
The ~b~I"r"~N means that access to the file is for ~I"r"~Neading ( as
opposed to "w"riting or "a"ppending ).
.WR20C1
.R20C1
How do we tell ~b~Ifopen()~N the name of the file which we want to open...?
~b~Iname[]~N is an ARRAY of pointers, each pointing to a string.
.WR20C1
.R20C1
When we typed ~Iread letter~N, we passed to our ~b~Imain()~N program the
two strings ~Iread~N and ~Iletter~N. Now the pointer ~b~Iname[0]~N points to
~Iread~N and ~b~Iname[1]~N will point to the string ~Iletter~N ...
.WR20C1
.R20C1
...so we give the name of the file to ~b~Ifopen()~N, namely ~b~Iname[1]~N.
.K10,60
of course
.WR14C1
6 ~b~I fp=fopen(name[1],"r"); ~N
~V7 if (fp == NULL) { ~N
~V8 printf("\nSorry, can't open %s",name[1]); ~N
~V9 exit(0); ~N
.R20C1
.R20C1
The function ~b~Ifopen()~N is clever. If it can't open the file for some
reason (beer-on-the-disk) it ~Ireturn~Ns a NULL pointer.
We check, in Line 7, IF ~b~Ifp~N is EQUAL ( ~b~I==~N ) to ~b~INULL~N.
If so, we apologize ( Line 8 ) and ~b~Iexit~N the program ( Line 9 ).
.WR15C1
7 ~b~I if (fp == NULL) { ~N
8 ~b~I printf("\nSorry, can't open %s",~Vname[1]~N~b~I); ~N
9 ~b~I ~Vexit(0)~N~b~I; ~N
.R20C1
.R20C1
In Line 8 we ~b~Iprintf~N ~b~Iname[1]~N as a ~b~I%s~Ntring (of course).
The ~b~Iexit(0)~N (in Line 9 ) is new. It will exit our ~b~Imain()~N
and return you to the operating system (e.g. DOS ). ( It also ~Icloses~N
any open files ).
.WR20C1
.R20C1
NOTE: We gave to ~b~Ifopen()~N the string which ~b~Iname[1]~N points to,
and we also give it to ~b~Iprintf()~N (namely ~Iletter~N, in this
example).
.WNK10,32
phew!!
.WN
1 ~b~Imain(number,name) ~N
2 ~b~Iint number; ~N
3 ~b~Ichar *name[] ~N
4 ~b~I{ ~N
5 ~b~I FILE *fp, *fopen(); ~N
6 ~b~I fp=fopen(*name[1],"r"); ~N
7 ~b~I if (fp == NULL) { ~N
8 ~b~I printf("\nSorry, can't open %f",name[1]); ~N
9 ~b~I exit(0); ~N
10 ~b~I } ~N
11 ~b~I /* now we read from the file */ ~N
~V12 char c; ~N
~V13 while ( (c=getc(fp)) != EOF ) ~N
~V14 printf("%c",c); ~N
~V15 } ~N
.w
Now we declare a ~b~Ichar c~N ( Line 12 ) and continually ~b~Igetc()~N
~Ifrom the file which we opened~N ( in Line 6 ). Reference to this open
file is via its ~Ifile pointer~N ~b~Ifp~N ... hence we used ~b~Igetc(~Ffp~N~b~I)~N
to tell the ~b~Igetc()~N function which file to ~b~Iget~N the next
~b~Ic~Nhar from. ( ~b~Igetc()~N is in ~b~Istdio.h~N ).
.WR17C1
.R17C1
We (in this example) ~b~Iprintf()~N the ~%c~Nhar to the screen, ~b~Iwhile~N
~b~Ic~N is NOT EQUAL ( ~b~I!=~N ) to the ~b~IEOF~N character (which ~Ishould~N
indicate ~IE~Nnd ~IO~Nf ~IF~Nile ..if all goes well ..).
.WR17C1
.R17C1
.Q
How many errors in the above program ?
4
.R21C1
Line 3 needs a ~Isemi-colon~N!
.R3C1
3 ~b~Ichar *name[] ~Ftch! tch!~N~b~I~N
.WR21C1
Line 6 should give ~b~Iname[1]~N to ~b~Ifopen()~N, NOT ~b~I*name[1]~N
.R6C1
6 ~b~I fp=fopen(~F*~N~b~Iname[1],"r"); ~N
.WR21C1
Line 8 needs a ~b~I%s~Ntring format, NOT a ~b~I%f~Nloat !
.R8C1
8 ~b~I printf("\nSorry, can't open %~Ff~N~b~I",name[1]); ~N
.WR21C1
....and we need to ~b~I#include <stdio.h>~N
.K10,60
as Line 0
.WNK10,32
funtastic!
.WN
0 ~b~i#include <stdio.h> ~N
1 ~b~Imain(number,name) ~N
2 ~b~Iint number; ~N
3 ~b~Ichar *name[]; ~N
4 ~b~I{ ~N
5 ~b~I FILE *fp, *fopen(); ~N
6 ~b~I fp=fopen(name[1],"r"); ~N
7 ~b~I if (fp == NULL) { ~N
8 ~b~I printf("\nSorry, can't open %s",name[1]); ~N
9 ~b~I exit(0); ~N
10 ~b~I } ~N
Let's begin with this introduction ( ...same as before, but with ~Cbugs~N
removed!) and continue with a program which will read a file on disk
and count the number of times the letters 'a' through 'z' occur.
We'll compile/link the program under the name ~Icount~N and we will
type ~Icount sam~N to count the letters 'a'-'z' in a file called ~Isam~N.
.K19,32
how nice!
.WN
.T
counting from 'a to 'z'
.WN
~Vchar c; /* characters from file */~N
~Vint n[26], i; /* count for each letter*/~N
~b~Ifor (i=0;i<26;i++) n[i]=0; /* counts set to zero */~N
~b~Ifor (i=0;i<1000;i++) { /* go thru 1000 chars */~N
~b~I c=getc(fp); /* get a char from file */~N
~b~I if (c>='a'&& c<='z') { /* is it between a & z? */~N
~b~I n[c-'a']++; /* yes? increment count */~N
~b~I } /* end of if */~N
~b~I} /* end of for */~N
~b~Iprintf("\nIn the file %s : ",name[1]); /* print file name */~N
~b~Ifor (i=0;i<=24;i+=2) /* go thru alphabet */~N
~b~I printf("\n%c%s%3d%s : %c%s%3d%s", /* a format string */~N
~b~I 'a'+i, " occurs ",n[i], " times", /* print letter and */~N
~b~I 'a'+i+1," occurs ",n[i+1]," times");/* count for each letter*/~N
~b~I} /* end of main() */~N
.R17C1
Here we declare some stuff, notably the ARRAY ~b~In[26]~N which will hold
the number of times each of the 26 letters from 'a' to 'z' occur ...hence
it's an ~b~Iint~Neger ARRAY.
.WR1C1
~b~Ichar c; /* characters from file */~N
~b~Iint n[26], i; /* count for each letter*/~N
~Vfor (i=0;i<26;i++) n[i]=0; /* counts set to zero */~N
.R17C1
.R17C1
Here we set all the integers in the ARRAY to 0 ..you can never tell what
garbage might be in those memory locations!
.WR3C1
~b~Ifor (i=0;i<26;i++) n[i]=0; /* counts set to zero */~N
~Vfor (i=0;i<1000;i++) { /* go thru 1000 chars */~N
~V c=getc(fp); /* get a char from file */~N
~V if (c>='a'&& c<='z') { /* is it between a & z? */~N
.R17C1
.R17C1
Now we go through 1000 characters in the file and ~b~Igetc(fp)~N each one,
giving the file ~r~Ipointer~N ~b~Ifp~N to the library function ~b~Igetc()~N
and getting in ~Ireturn~N a ~b~Ichar~Nacter ~b~Ic~N which we check to see
if it's GREATER or EQUAL to the character ~b~I'a'~N AND ( ~b~I&&~N )
if it's LESS or EQUAL to the character ~b~I'z'~N.
.WR4C1
~b~Ifor (i=0;i<1000;i++) { /* go thru 1000 chars */~N
~b~I c=getc(fp); /* get a char from file */~N
~b~I if (c>='a'&& c<='z') { /* is it between a & z? */~N
~V n[c-'a']++; /* yes? increment count */~N
.R17C1
.R17C1
~Iif~N we get a character between 'a' and 'z' we increment the appropriate
member of our ARRAY ~b~In[]~N ....but how ??
.WR17C1
.R17C1
Now ~b~Ic~N is a (single) byte in memory, containing the 'value' of the
~b~Ic~Nharacter. This 'value' would be ~I65~N for an 'a' and ~I66~N
for a 'b', etc. (in ASCII). If the 'value' of the ~b~Ic~Nharacter returned
by ~b~Igetc()~N were ~I69~N (for example) then we must have an 'e' and we
would increment ~b~In[4]~N (which holds the number of times an 'e' occurs).
BUT, the 'value' of ~b~Ic~N minus the 'value' of 'a' is ~I69-65=4~N ...
...that is, ~b~Ic-'a'~N has the 'value' ~I4~N ...
SO we increment ~b~In[c-'a']~N ( which WILL increment n[4] ).
.WR7C1
~b~I n[c-'a']++; /* yes? increment count */~N
~V } /* end of if */~N
~V} /* end of for */~N
.R17C1
.R17C1
Now we end the ~Iif~N and the ~Ifor~N.
The numbers in the ARRAY ~b~In[]~N will give us what we want.
.WR8C1
~b~I } /* end of if */~N
~b~I} /* end of for */~N
~Vprintf("\nIn the file %s : ",name[1]); /* print file name */~N
.R17C1
.R17C1
Now (just for kicks) we print the ~b~Iname[1]~N of the file we opened...
.WR10C1
~b~Iprintf("\nIn the file %s : ",name[1]); /* print file name */~N
~Vfor (i=0;i<=24;i+=2) /* go thru alphabet */~N
.R17C1
.R17C1
Now we go through all 26 letters ( and all 26 'counts' stored in the
ARRAY ~b~In[]~N ) but two-at-a-time ( which is why we used ~b~Ii+=2~N
which adds ~I2~N to ~b~Ii~N each time through the ~Ifor~N loop.
.WR11C1
~b~Ifor (i=0;i<=24;i+=2) /* go thru alphabet */~N
~V printf("\n%c%s%3d%s : %c%s%3d%s", /* a format string */~N
~V 'a'+i, " occurs ",n[i], " times", /* print letter and */~N
~V 'a'+i+1," occurs ",n[i+1]," times");/* count for each letter*/~N
.R17C1
.R17C1
Now we print each letter as a ~n~I%c~Nharacter, giving to ~b~Iprintf()~N
the numbers 'a'+0, then 'a'+1, then 'a'+2, etc. (meaning, the ASCII
'values' ~I65~N then ~I66~N then ~I67~N etc.).
We also print ~b~In[0]~N, ~b~In[1]~N, ~b~In[2]~N etc. which are the
various 'counts' (for an 'a', 'b', 'c', etc.)
.WR12C1
~V printf("\n~N~b~I%c%s%3d%s~N~V:%c%s%3d%s", /* a format string */~N
.R17C1
.R17C1
Note that we choose (for variety) to specify a format which says to print:
First a ~b~I%c~Nharacter then a ~b~I%s~Ntring~N then a ~b~I%d~Necimal (field width ~I3~N)
then another ~b~I%s~Ntring~N.
.WR12C1
~V printf("\n~F%c%s%3d%s~N~V:~F%c%s%3d%s~N~V", /* a format string */~N
.R17C1
.R17C1
...and we do this TWICE (before we go to a ~b~I\n~Newline) which is why
we increased ~b~Ii~N by ~I2~N each time through the loop (as in ~b~Ii+=2~N).
.WR12C1
~b~I printf("\n%c%s%3d%s : %c%s%3d%s", /* a format string */~N
~V 'a'+i, " occurs ",n[i], " times", /* print letter and */~N
~b~I 'a'+i+1," occurs ",n[i+1]," times");/* count for each letter*/~N
.R17C1
.R17C1
This is one of the TWO printouts on each line of the screen:
First the ~b~I%c~Nharacter (as ~b~I'a'+i~N) then the ~b~I%s~Ntring (" occurs ")
then the ~b~I%d~Necimal ( ~b~In[i]~N ) with field width ~I3~N, then the ~b~I%s~Ntring
" times".
.WR13C1
~b~I 'a'+i, " occurs ",n[i], " times", /* print letter and */~N
~b~I 'a'+i+1," occurs ",n[i+1]," times");/* count for each letter*/~N
~V} /* end of main() */~N
.R17C1
.R17C1
..and this (would you believe) is the end of ~b~Imain()~N !
(We really ~Ishould~N provide that 4-space indent!)
.WK2,32
printout??
.WN
For the program we just wrote (called ~Icount.c~N, before it was compiled
and linked) we get:
~IA>count count.c~N
~r~IIn the file count.c : ~N
~r~Ia occurs 18 times : b occurs 3 times~N
~r~Ic occurs 19 times : d occurs 4 times~N
~r~Ie occurs 32 times : f occurs 22 times~N
~r~Ig occurs 3 times : h occurs 9 times~N
~r~Ii occurs 25 times : j occurs 0 times~N
~r~Ik occurs 0 times : l occurs 5 times~N
~r~Im occurs 11 times : n occurs 31 times~N
~r~Io occurs 19 times : p occurs 9 times~N
~r~Iq occurs 0 times : r occurs 23 times~N
~r~Is occurs 9 times : t occurs 22 times~N
~r~Iu occurs 7 times : v occurs 0 times~N
~r~Iw occurs 1 times : x occurs 1 times~N
~r~Iy occurs 2 times : z occurs 3 times~N
.K10,60
pretty!
.WN
~b~Ifor (i=0;i<=24;i+=2) /* go thru alphabet */~N
~b~I printf("\n%c%s%3d%s : %c%s%3d%s", /* a format string */~N
~b~I 'a'+i, " occurs ",n[i], " times", /* print letter and */~N
~b~I 'a'+i+1," occurs ",n[i+1]," times");/* count for each letter*/~N
The mechanism for printing two-to-a-line is awkward (!)
We really want to let ~b~Ii~N go from ~b~I0~N to ~b~I25~N and, each time ~b~Ii~N
is exactly divisible by ~I2~N, printf a ~b~I\n~Newline.
This 'exact divisibility' can be tested by dividing ~b~I~N by ~I2~N and checking
if the remainder is zero.
To get the remainder, when ~b~Ii~N is divided by ~I2~N, we use: ~b~Ii%2~N
(which gives this remainder!).
.W
~b~Ifor (i=0;~Vi<26;i++~N~b~I) /* go thru alphabet */~N
~V if (i%~F2~N~V==0) printf("\n"); /* newline if i%2==0 */~N
~b~I printf("%c%s%3d%s", /* a format string */~N
~b~I 'a'+i, " occurs ",n[i], " times"); /* print letter count */~N
.R22C1
...much nicer code, since we can easily print 3 or 4 to a line by changing
the number ~V~F2~N.
.WN
.T
That's all folks!
.K16,32
au revoir!
.q