csv

commit eb5cc48c716d078349eda821bab263af80757fe3
parent d666838b2ad0776a1e2bb9fef8e2e8ef6de60278
Author: aabacchus <ben@bvnf.space>
Date:   Sat,  2 Oct 2021 16:14:32 +0100

move data reading into its own function

Diffstat:
M
csv.c
 | 
130
+++++++++++++++++++++++++++++++++++++++++++++++--------------------------------

1 file changed, 77 insertions(+), 53 deletions(-)
diff --git a/csv.c b/csv.c
@@ -34,6 +34,74 @@ read_header(FILE *f, char ***header) {
 	return ncols;
 }
 
+/* read_data reads all the comma-separated values from f into *datap,
+ * and returns the number of rows.
+ * This function mallocs the number of rows in *header[0..n-1]
+ * -1 is returned on error.
+ */
+int
+read_data(FILE *f, float ***datap, int ncols) {
+	float **data = *datap;
+	int nrows = 0;
+	char *n = malloc(MAX_LINE);
+	if (n == NULL) {
+		perror("malloc");
+		return -1;
+	}
+	char *np = n;
+	data[0] = malloc(sizeof (float) * ncols);
+	if (data[0] == NULL) {
+		perror("malloc");
+		free(n);
+		return -1;
+	}
+	int c, x, y;
+	x = y = 0;
+	while ((c = fgetc(f)) != EOF) {
+		if (c == ',' || c == '\n') {
+			if (n == np) {
+				/* nothing in this column */
+				fprintf(stderr, "warning: empty cell at row %d, col %d\n", y + 1, x + 1);
+				data[y][x] = 0.;
+			} else {
+				*np++ = '\0';
+				data[y][x] = (float)atof(n);
+				np = n;
+			}
+			x++;
+			if (c == '\n') {
+				nrows++;
+				y++;
+				x = 0;
+				float **tmp = realloc(data, sizeof (float *) * (nrows + 1));
+				if (tmp == NULL) {
+					perror("malloc");
+					goto cleanup;
+				}
+				data = tmp;
+				data[y] = malloc(sizeof (float) * ncols);
+				if (data[y] == NULL) {
+					nrows--;
+					perror("malloc");
+					goto cleanup;
+				}
+				continue;
+			}
+			continue;
+		}
+		*np++ = c;
+	}
+	free(n);
+	*datap = data;
+	return nrows;
+cleanup:
+	free(n);
+	for (int i = 0; i < nrows; y++)
+		free(data[i]);
+	free(data);
+	return -1;
+}
+
 int
 main(int argc, char **argv) {
 	int c;
@@ -89,54 +157,16 @@ read from stdin\n"
 	}
 
 	/* read data */
-	char *n = malloc(MAX_LINE);
-	char *np = n;
-	int nrows = 0;
-	float **data;
-	data = malloc(sizeof (float *) * (nrows + 1));
+	float **data = malloc(sizeof (float *) * 1);
 	if (data == NULL) {
 		perror("malloc");
-		goto fail;
-	}
-	data[0] = malloc(sizeof (float) * ncols);
-	if (data[0] == NULL) {
-		perror("malloc");
-		goto fail;
+		return 1;
 	}
-	np = n;
-	int x, y;
-	x = y = 0;
-	while ((c = fgetc(f)) != EOF) {
-		if (c == ',' || c == '\n') {
-			if (n == np) {
-				/* nothing in this column */
-				fprintf(stderr, "warning: empty cell at row %d, col %d\n", y + 1, x + 1);
-				data[y][x] = 0.;
-			} else {
-				*np++ = '\0';
-				data[y][x] = (float)atof(n);
-				np = n;
-			}
-			x++;
-			if (c == '\n') {
-				nrows++;
-				y++;
-				x = 0;
-				data = realloc(data, sizeof (float *) * (nrows + 1));
-				if (data == NULL) {
-					perror("malloc");
-					goto fail;
-				}
-				data[y] = malloc(sizeof (float) * ncols);
-				if (data[y] == NULL) {
-					perror("malloc");
-					goto fail;
-				}
-				continue;
-			}
-			continue;
-		}
-		*np++ = c;
+	int nrows = read_data(f, &data, ncols);
+	if (nrows == -1) {
+		free(data);
+		fclose(f);
+		return 1;
 	}
 	for (int i = 0; i < nrows; i++) {
 		for (int j = 0; j < ncols; j++)
@@ -145,15 +175,9 @@ read from stdin\n"
 	}
 	fprintf(stderr, "table has %d rows of %d columns\n", nrows, ncols);
 
-	free(n);
+	for (int i = 0; i < nrows; i++)
+		free(data[i]);
 	free(data);
-
-
 	fclose(f);
 	return 0;
-fail:
-	fclose(f);
-	free(n);
-	free(data);
-	return 1;
 }