- When processing regular characters (final else clause) you read additional input instead of operating over
text
, i.e. instead of:
scanf("%c", &document[parano][sentno][wordno][charno]);
printf("%c\n", document[parano][sentno][wordno][charno]);
charno++;
you want to do:
document[parano][sentno][wordno][charno++] = text[i];
By the time we hit text[i] == ' '
for the first time the value of ***document
is “3\n1 2\n2\n” but you wanted it to be “Learning”.
-
(not fixed) When processing a word (`text[i] == ‘ ‘) you expand current word, yet, you hard-code 1000 when you allocate it initially so this doesn’t make sense. Be consistent.
-
parano
,sentno
,wordno
,charno
is indices but same suggest they are counts. It’s not wrong just confusing and why you have towordno + 2
when relloc’ing. -
Terminate words with `\0′.
-
When you process ‘ ‘ you add another word which you may or may not need which is fine. But when process a ‘.’ you look ahead to the following letter is a
\n
or not. Be consistent. If you look ahead then you need to check thati + 1 < len
, and it’s fragile, say, there is a stray space before the\n
. -
(not fixed) Memory leaks. As the size of the sub-elements (paragraph, sentences and words) are private implementation details of
get_document()
you will have refactor the code to make those available. I suggest:
struct document {
char ****document;
size_t paragraphs;
size_t sentences;
size_t words;
}
-
(not fixed) Deduplicate. Similar to the print functions, create a allocation function per type.
-
(not fixed, really)
get_input_text()
. You split the into paragraphs then concatenate everything again into a local variable then copy it into a dynamically allocated variable:
char *str = malloc(BUFFER_LEN);
for(size_t i = 0, l = 0; l < lines && i + 1 < BUFFER_LEN; i += strlen(str + i)) {
int rv = fgets(str + i, BUFFER_LEN - i, stdin);
if(!rv) {
// handle error
break;
}
}
return s;
- (not fixed) Separate i/o from processing. This simplifies testing and it makes it easier to figure out what is going on. In
main(), you read a query type then 1 to 3 numbers.
scanf()` tells you how many items where read so you simply do. As you don’t use the kth_ functions for anything else just combine then with print_ funci
int n = scanf("%d %d %d %d", &type, &p, &s, &w);
switch(type) {
case 1:
if(n != 2) {
// handle error
}
print_paragraph(document, p);
break;
...
}
-
(not fixed) Add error checks for the remaining
malloc()
,strdup()
etc. -
Don’t hard-code magic values (1000). I introduced the constant
WORD_LEN
but you also have MAX_CHARACTERS which is kinda the same thing. -
(not fixed) Consider using
char *s = strpbrk(text + i, " .\n")
to copy a word at a time. It will simplify\0
handling, and likely to be faster than walking thetext
a byte at a time, i.e.i += s - text + i
, just handle thes == NULL
special case.
valgrind is now happy other than leaks (see above):
#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define MAX_CHARACTERS 1005
#define MAX_PARAGRAPHS 5
#define WORD_LEN 1000
char *kth_word_in_mth_sentence_of_nth_paragraph(char ****document, int p, int s, int w) {
return document[p - 1][s - 1][w - 1];
}
char **kth_sentence_in_mth_paragraph(char ****document, int p, int s) {
return document[p - 1][s - 1];
}
char ***kth_paragraph(char ****document, int p) {
return document[p - 1];
}
char ****get_document(char* text) {
char ****document = malloc(sizeof ***document);
*document = malloc(sizeof **document);
**document = malloc(sizeof *document);
***document = malloc(WORD_LEN);
/* declare some numbers as iterators for the words, sentences,etc.*/
int parano = 0;
int sentno = 0;
int wordno = 0;
int charno = 0;
/* now iterate over the text filling and expanding the document at the same time*/
/*----------------------------------------------------------------------------------------------------*/
/* feading data in those spaces*/
size_t len = strlen(text);
for (size_t i = 0; i < len; i++) {
switch(text[i]) {
case ' ': {
document[parano][sentno][wordno][charno] = '\0';
wordno++;
char **words = realloc(
document[parano][sentno],
(wordno + 1) * sizeof *document[parano][sentno]
);
if(!words) {
printf("realloc of words failed\n");
exit(1);
}
document[parano][sentno] = words;
document[parano][sentno][wordno] = malloc(WORD_LEN);
charno = 0;
break;
}
case '.': {
document[parano][sentno][wordno][charno] = '\0';
sentno++;
char ***sentences = realloc(
document[parano],
(sentno + 1) * sizeof *document[parano]
);
if(!sentences) {
printf("realloc of sentences failed\n");
exit(1);
}
document[parano] = sentences;
document[parano][sentno] = malloc(sizeof **document);
wordno = 0;
document[parano][sentno][wordno] = malloc(WORD_LEN);
charno = 0;
break;
}
case '\n': {
document[parano][sentno][wordno][charno] = '\0';
parano++;
char ****paragraphs = realloc(
document,
(parano + 1) * sizeof *document
);
if(!paragraphs) {
printf("realloc of paragraphs failed\n");
exit(1);
}
document = paragraphs;
document[parano] = malloc(sizeof ***document);
sentno = 0;
document[parano][sentno] = malloc(sizeof **document);
wordno = 0;
document[parano][sentno][wordno] = malloc(WORD_LEN);
charno = 0;
break;
}
default: // character
document[parano][sentno][wordno][charno++] = text[i];
}
}
return document;
}
char *get_input_text() {
int paragraph_count;
scanf("%d", ¶graph_count);
char p[MAX_PARAGRAPHS][MAX_CHARACTERS];
char doc[MAX_CHARACTERS];
memset(doc, 0, sizeof doc);
getchar();
for (int i = 0; i < paragraph_count; i++) {
scanf("%[^\n]%*c", p[i]);
strcat(doc, p[i]);
if (i != paragraph_count - 1)
strcat(doc, "\n");
}
return strdup(doc);
}
void print_word(char *word) {
printf("%s", word);
}
void print_sentence(char **sentence) {
int word_count;
scanf("%d", &word_count);
for(int i = 0; i < word_count; i++){
print_word(sentence[i]);
if(i + 1 != word_count)
printf(" ");
}
}
void print_paragraph(char ***paragraph) {
int sentence_count;
scanf("%d", &sentence_count);
for (int i = 0; i < sentence_count; i++) {
print_sentence(paragraph[i]);
printf(".");
}
}
int main() {
char *text = get_input_text();
char ****document = get_document(text);
int q;
scanf("%d", &q);
while (q--) {
int type;
scanf("%d", &type);
switch(type) {
case 1: {
int p;
scanf("%d", &p);
print_paragraph(kth_paragraph(document, p));
break;
}
case 2: {
int p, s;
scanf("%d %d", &p, &s);
print_sentence(kth_sentence_in_mth_paragraph(document, p, s));
break;
}
case 3: {
int p, s, w;
scanf("%d %d %d", &p, &s, &w);
print_word(kth_word_in_mth_sentence_of_nth_paragraph(document, p, s, w));
break;
}
default:
printf("error\n");
}
printf("\n");
}
free(text);
}
Output as expected:
Learning pointers is more fun.It is good to have pointers.
Learning C is fun
Learning
Btw, a whole different way of solving this is problem is keep the original input (text
) then write functions to directly extract a paragraph, sentence or word from that string. If you input is huge then create an index, say, (paragraph
, sentence
, word
) to &text[i]
.
1
solved i have been on this question for quaring the document for two days straight, and it is not working. don’t want to cheat, can you point the problem