File I/O — Personalized Mail Merge: 25-Comp-A4 Program Design and Data Structures, December 2016, Q4

Question 4: File I/O — Personalized Mail Merge (20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Given. A template text file containing zero or more occurrences of the literal marker #N#, and a recipient name typed at the keyboard.

Find. A copy of the template, written to a second file, with every occurrence of #N# replaced by the typed name.

Approach. Stream the template character by character; whenever the next three characters match #N# exactly, write the name instead and skip past all three marker characters, otherwise copy the character through unchanged.

Write the program. A small 3-character lookahead buffer (implemented here with ungetc, so no separate buffering scheme is needed) lets the marker be recognised without reading the whole file into memory.

#include <stdio.h>
#include <string.h>

int main(void)
{
    FILE *in, *out;
    char name[80];
    int c1, c2, c3;

    in = fopen("letter_template.txt", "r");
    out = fopen("letter_out.txt", "w");
    printf("Enter recipient's first name: ");
    fgets(name, sizeof name, stdin);
    name[strcspn(name, "\n")] = '\0';        /* strip trailing newline */

    c1 = fgetc(in);
    while (c1 != EOF) {
        if (c1 == '#') {
            c2 = fgetc(in);
            if (c2 == 'N') {
                c3 = fgetc(in);
                if (c3 == '#') {
                    fputs(name, out);         /* marker matched: substitute */
                    c1 = fgetc(in);
                    continue;
                }
                fputc(c1, out); fputc(c2, out);
                if (c3 != EOF) ungetc(c3, in);
                c1 = fgetc(in);
                continue;
            }
            fputc(c1, out);
            if (c2 != EOF) ungetc(c2, in);
            c1 = fgetc(in);
            continue;
        }
        fputc(c1, out);
        c1 = fgetc(in);
    }
    fclose(in);
    fclose(out);
    return 0;
}

Confirm on a short worked example. Template "Dear #N#, welcome #N#!" with name Chris streams through unchanged up to the first #, matches the full #N# marker there, substitutes Chris, continues copying ", welcome " unchanged, matches the marker a second time, and substitutes again — confirming the "may occur more than once" requirement is met without any assumption that there is exactly one marker. $$\boxed{\text{"Dear \#N\#, welcome \#N\#!"} \to \text{"Dear Chris, welcome Chris!"}}$$

Question 4 — results
Template	Output (name = "Chris")
Dear #N#, welcome #N#!	Dear Chris, welcome Chris!

Concept & Theory

Topic. Streaming multi-character pattern replacement with bounded lookahead, and dual-stream (input file + output file) text processing.

Key relations. A literal $k$-character marker can be matched by a streaming scan using at most $k-1$ characters of pushback/lookahead, without the naive-matching worst case that full string-search algorithms (KMP, Boyer-Moore) are built to avoid — here $k=3$ is small enough that naive character-by-character matching with ungetc is simplest and entirely adequate.

Why this works. Because the marker is a fixed literal (not a variable pattern), the scan only needs to remember "how much of #N# have I matched so far" and, on a mismatch, put back every character it provisionally consumed beyond the first so the next comparison starts from the correct position — this is exactly what ungetc provides for a 1-character-of-file-buffer pushback, chained twice here for a 2-character worst case.

Common pitfall. Reading the marker with three separate fgetc calls and, on a partial match (e.g. #N followed by a non-#), forgetting to push the non-matching character back — which silently drops one input character from the output; and not stripping the newline from the keyboard-entered name, which then embeds a stray line break into every substitution.

Source. Kernighan & Ritchie, The C Programming Language, 2nd ed., §7.4–7.5 (file I/O, character streams); Weiss, Data Structures and Algorithm Analysis in C, 2nd ed., ch. 1 (string processing).

Question 4 of 9: File I/O — Personalized Mail Merge

Question 4: File I/O — Personalized Mail Merge (20 marks)