freq

commit ce8461f31a92eb9a5988450303db8337a8864406
Author: phoebos <ben@bvnf.space>
Date:   Mon, 21 Mar 2022 15:58:13 +0000

initial

Diffstat:
A
.gitignore
 | 
1
+
A
Makefile
 | 
6
++++++
A
freq.c
 | 
123
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

3 files changed, 130 insertions(+), 0 deletions(-)
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1 @@
+freq
diff --git a/Makefile b/Makefile
@@ -0,0 +1,6 @@
+.POSIX:
+
+XCFLAGS = $(CFLAGS) -Wall -Wextra -pedantic -Og -g -D_XOPEN_SOURCE=700
+
+freq: freq.c
+	$(CC) $(XCFLAGS) freq.c -o $@
diff --git a/freq.c b/freq.c
@@ -0,0 +1,123 @@
+#include <ctype.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+void freq(FILE *f, int field);
+
+struct count {
+	char *name;
+	int freq;
+};
+
+struct count *cs;
+size_t cs_len = 0;
+
+size_t
+get_index(char *name) {
+	for (size_t i = 0; i < cs_len; i++) {
+		if (strcmp(cs[i].name, name) == 0)
+			return i;
+	}
+	return -1;
+}
+
+/* can only increase freqs */
+void
+update_record(char *name) {
+	int i = get_index(name);
+	if (i == -1) {
+		/* create record */
+		cs_len++;
+		cs = realloc(cs, sizeof(struct count) * cs_len);
+		if (cs == NULL) {
+			perror("realloc");
+			exit(1);
+		}
+		cs[cs_len - 1].name = strdup(name);
+		cs[cs_len - 1].name = 0;
+	} else {
+		/* add 1 to freq */
+		cs[i].freq++;
+	}
+}
+
+void
+print_freqs(void) {
+	for (size_t i = 0; i < cs_len; i++)
+		printf("%d %s\n", cs[i].freq, cs[i].name);
+}
+
+int
+main (int argc, char **argv) {
+	int c, field = 0;
+	while ((c = getopt(argc, argv, ":k:")) != -1) {
+		switch (c) {
+			case 'k':
+				field = atoi(optarg);
+				break;
+			case '?':
+				fprintf(stderr, "usage: %s [-k fieldnum] [file]\n", argv[0]);
+				return 1;
+		}
+	}
+	argc -= optind;
+	argv += optind;
+
+	FILE *f = stdin;
+	if (*argv) {
+		if (strcmp(*argv, "-") != 0) {
+			f = fopen(*argv, "r");
+			if (f == NULL) {
+				perror(*argv);
+				return 1;
+			}
+		}
+	}
+
+	freq(f, field);
+
+	if (f != stdin)
+		fclose(f);
+	return 0;
+}
+
+void
+freq(FILE *f, int field) {
+	char *line = NULL;
+	size_t linelen = 0;
+	ssize_t n;
+	int cur_field = 1;
+	int in_word = 0;
+
+	while ((n = getline(&line, &linelen, f)) != -1) {
+		ssize_t i = 0;
+		char *field_start = line;
+		/* eat up whitespace at front */
+		while (isspace(line[i++]));
+		for (; i < n; i++) {
+			if (isspace(line[i])) {
+				if (in_word) {
+					if (cur_field == field) {
+						line[i] = '\0';
+						goto field_found;
+					}
+					cur_field++;
+					in_word = 0;
+				}
+			} else {
+				if (!in_word)
+					field_start = line + i;
+				in_word = 1;
+			}
+		}
+field_found:
+		fprintf(stderr, "found '%s'\n", field_start);
+		update_record(field_start);
+
+	}
+	free(line);
+
+	print_freqs();
+}