\documentstyle{article} \setlength{\oddsidemargin}{0in} \setlength{\evensidemargin}{0in} \setlength{\topmargin}{0.0in} \setlength{\headheight}{0in} \setlength{\headsep}{0in} \setlength{\textwidth}{6.5in} \setlength{\textheight}{9.6in} \setlength{\parindent}{0in} \setlength{\parskip}{0.12in} \begin{document} \newcommand{\bfs}[1]{{\bf #1}} \begin{center} {\LARGE\bf Introduction to C File Access } \end{center} \bigskip The program below will serve as a first introduction to reading and writing files (directly, not through the $<$ and $>$ redirection methods) in C. It will also introduce a few new C constructs. The sample program first prompts the user to specify the names of input and output files. The program will then read the input file, and search for any single-digit, nonzero words; it will then produce an identical output file, except that the single-digit words have been changed from arabic to roman numerals. Here is a \bfs{script} record of a sample run: \begin{verbatim} 1 Script started on Wed Apr 15 23:52:27 1992 2 heather% cat x 3 4 She has 7 sisters, 5 2 brothers, 6 nieces and 3 6 nephews. 7 8 heather% a.out 9 enter name of input file 10 x 11 enter name of output file 12 y 13 heather% cat y 14 15 She has VII sisters, 16 II brothers, VI nieces and III 17 nephews. 18 19 heather% e 20 heather% 21 script done on Wed Apr 15 23:52:46 1992 \end{verbatim} Here is the program. Some remarks will follow it. \begin{verbatim} 1 2 3 /* converts single-digit, nonzero arabic numbers to roman numbers 4 in a given text file */ 5 6 7 /* WARNING: for demonstration purposes only; not completely general */ 8 9 10 #include 11 12 13 char InFile[1000], /* the entire input file */ 14 InFileName[20], /* name of the input file */ 15 OutFileName[20], /* name of the output file */ 16 Word[25]; /* current word */ 17 18 19 int InFileSize, /* number of characters in the input file */ 20 CurrInFilePlace, /* current scanning place in the input file; 21 used when we are alternating looking for 22 blanks and words */ 23 WordLength; /* length of the current word */ 24 25 26 FILE *InFilePtr, /* pointer to input file */ 27 *OutFilePtr; /* pointer to output file */ 28 29 30 ReadInFile() /* read the input file into the array InFile */ 31 32 { char C; 33 34 InFileSize = 0; 35 while (fscanf(InFilePtr,"%c",&C) != -1) InFile[InFileSize++] = C; 36 } 37 38 39 PrintFile() /* for debugging purposes only */ 40 41 { int I; 42 43 for (I = 0; I < InFileSize; I++) printf("%c",InFile[I]); 44 } 45 46 47 GetFile() 48 49 { printf("enter name of input file\n"); 50 scanf("%s",InFileName); 51 printf("enter name of output file\n"); 52 scanf("%s",OutFileName); 53 InFilePtr = fopen(InFileName,"r"); 54 OutFilePtr = fopen(OutFileName,"w"); 55 ReadInFile(); 56 } 57 58 59 WriteBlanksAndEOLs() /* starting at the current position in the 60 input file, keep scanning until get a 61 "real" character, i.e. not a blank or 62 end-of-line; meanwhile, copy all the 63 blanks and EOLs to the output file */ 64 65 { char C; 66 67 while (1) { 68 C = InFile[CurrInFilePlace]; 69 if (C == ' ' || C == '\n') { 70 fprintf(OutFilePtr,"%c",C); 71 CurrInFilePlace++; 72 } 73 else break; 74 } 75 } 76 77 78 int GetWord() /* scans this word until hit its end; copy the word 79 to the array Word; also, return 1 if successful 80 in getting a word, 0 otherwise (failed due to 81 reaching the end of the file) */ 82 83 { char C; 84 85 WordLength = 0; 86 87 while (1) { 88 /* if reach the end of the file, leave, reporting failure */ 89 if (CurrInFilePlace == InFileSize) return 0; 90 C = InFile[CurrInFilePlace]; 91 /* if hit the end of the word, leave, otherwise record the 92 current character in the array Word */ 93 if (C == ' ' || C == '\n') break; 94 else { 95 Word[WordLength++] = C; 96 CurrInFilePlace++; 97 } 98 } 99 100 return 1; 101 } 102 103 104 105 ConvertWord() /* the arabic-to-roman conversion is done here */ 106 107 { switch (Word[0]) { 108 case '1': Word[0] = 'I'; /* 1 = I */ 109 WordLength = 1; break; 110 case '2': Word[0] = Word[1] = 'I'; /* 2 = II */ 111 WordLength = 2; break; 112 case '3': Word[0] = Word[1] = Word[2] = 'I'; /* 3 = III */ 113 WordLength = 3; break; 114 case '4': Word[0] = 'I'; Word[1] = 'V'; /* 4 = IV */ 115 WordLength = 2; break; 116 case '5': Word[0] = 'V'; /* 5 = V */ 117 WordLength = 1; break; 118 case '6': Word[0] = 'V'; Word[1] = 'I'; /* 6 = VI */ 119 WordLength = 2; break; 120 case '7': Word[0] = 'V'; /* 7 = VII */ 121 Word[1] = Word[2] = 'I'; 122 WordLength = 3; break; 123 case '8': Word[0] = 'V'; /* 8 = VIII */ 124 Word[1] = Word[2] = Word[3] = 'I'; 125 WordLength = 4; break; 126 case '9': Word[0] = 'I'; Word[1] = 'X'; /* 9 = IX */ 127 WordLength = 2; 128 } 129 } 130 131 132 WriteWord() 133 134 { int I; 135 136 for (I = 0; I < WordLength; I++) fprintf(OutFilePtr,"%c",Word[I]); 137 } 138 139 140 main() 141 142 { GetFile(); 143 CurrInFilePlace = 0; 144 /* keep alternating this cycle: scan through blanks and end-of-line 145 characters, copying them to the output file */ 146 while (CurrInFilePlace < InFileSize) { 147 WriteBlanksAndEOLs(); 148 if (GetWord()) { 149 if (Word[0] > '0' && Word[0] <= '9' && WordLength == 1) 150 ConvertWord(); 151 WriteWord(); 152 } 153 } 154 } 155 156 \end{verbatim} \bfs{Analysis:} \bfs{Line 10.} This tells the C compiler (actually, the C preprocessor) to include C source code from a file stdio.h. Due to the presence of the `$<$ $>$', the file will be searched for in some standard directories, e.g. on most systems the directory /usr/include. (You should take a look at the file.) This file contains a bunch of definitions which are needed for the file manipulation we will be doing. \bfs{Lines 26-27.} Here we are declaring two \bfs{pointers} to a new type called `FILE'. That latter type is defined in the \#include file, and we will discuss pointers soon, but for the time being just take it on faith that we need them here. \bfs{Line 35.} Here we are using the \bfs{fscanf()} function, which is just like \bfs{scanf()}, except that it can be used on any file, not just the standard input. It has one more argument, which is the pointer to that file, in this case InFilePtr. Note that we keep looping until \bfs{fscanf()} tells us that we have reached the end of the file, which it does by returning a value of -1, just as with \bfs{scanf()} in earlier examples. \bfs{Lines 50 and 52.} Here we see the \%s format, for reading in character strings, just like \%c format reads in a character and \%d reads in an integer. Note that we could have used a \bfs{for} loop, with \%c format being used to read in the characters one at a time, but \%s is much more convenient. \bfs{Line 53.} Before accessing a file, we must ``open'' it first, using the \bfs{fopen()} function. Here we are opening the file whose name we have stored in the array InFileName, and we intend to read from it, which is why we have specified ``r''. The \bfs{fopen()} function then returns a pointer for us, which we assign to InFilePtr; this is how the pointer gets associated with the file. From that point on, all access to this file will be done via the pointer InFilePtr (as in Line 35).\footnote{Just as \bfs{scanf()} returns a value which can be used for error checking, so does \bfs{fopen()}. If the latter returns a 0 value, this means that it was not able to open the file.} \bfs{Line 54.} Same as Line 53, but with write access. \bfs{Line 69.} $||$ is the C symbol for \bfs{or}. \bfs{Line 70.} fprintf() is just like printf(), except for general files, not just the standard output. See remarks on \bfs{fscanf()} and \bfs{scanf()} above. \bfs{Lines 107-128.} A C \bfs{switch} statement is like a Pascal \bfs{case} statement. It is quite similar, except that in the C version, all cases past the one at hand will be done. For example, suppose Word[0] is `3'. Then if we didn't put in the \bfs{break} (Line 113), then the cases `4', `5', etc. would be done too! So the \bfs{break} is crucial; it says, ``Now leave this \bfs{switch} statement, and go to the first statement following the end of the \bfs{switch}.'' \end{document}