×

# KOL15A - Editorial

Author: Pankaj Jindal
Tester: Kevin Atienza
Editorialist: Kevin Atienza

Simple

# PREREQUISITES:

Type conversion, strings

# PROBLEM:

Given a string of alphanumeric characters, print the sum of the digits.

# EXPLANATION:

This is the easiest problem in the contest. Straightforward solutions will likely pass.

The idea is to iterate through the input string, filter out the non-digits, convert the digit characters into actual integers, and sum them. The following are a few sample implementations.

C:

#include <stdio.h>

char s[1111];
int main() {
int cases, cas, i, total;
scanf("%d", &cases);
for (cas = 0; cas < cases; cas++) {
scanf("%s", s);
total = 0;
for (i = 0; s[i]; i++) {
if ('0' <= s[i] && s[i] <= '9') total += s[i] - '0';
}
printf("%d\n", total);
}
return 0;
}


C++:

#include <iostream>
using namespace std;

int main() {
int cases;
cin >> cases;
for (int cas = 0; cas < cases; cas++) {
string s;
cin >> s;
int total = 0;
for (int i = 0; s[i]; i++) {
if ('0' <= s[i] && s[i] <= '9') total += s[i] - '0';
}
cout << total << endl;
}
return 0;
}


Java:

import java.util.Scanner;
public class Main {
public static void main(String args[]) {
Scanner sc = new Scanner(System.in);
int cases = sc.nextInt();
for (int cas = 0; cas < cases; cas++) {
String s = sc.next();
int total = 0;
for (int i = 0; i < s.length(); i++) {
char ch = s.charAt(i);
if ('0' <= ch && ch <= '9') {
total += ch - '0';
}
}
System.out.println(total);
}
}
}


# Implementation details

How to check if a character is a digit

All the above codes use the snippet '0' <= ch && ch <= '9' to determine if a character ch is a digit. This works because the char type in C, C++ and Java are implemented as a single byte representing the ASCII code of the character, and the digit characters '0' to '9' are consecutive ASCII codes. In fact, you can use a similar technique to check if ch is a lowercase character: 'a' <= ch && ch <= 'z'.

How to convert a character to an integer (if it is a digit)

Since characters are represented simply as bytes, we can perform arithmetic with them. This gives us a way to convert a digit character to a number. To convert a digit character ch to a number from $0$ to $9$, we simply use ch - '0'.

Note that a character being represented as a byte is not universal. In some other languages such as Python or Ruby, characters are actually represented as strings of length one, and although comparison between chars still work, you can't necessarily perform arithmetic between chars. For example, in Python, 'l' - 'a' would be an error. But usually there are still ways to get the ASCII value of a character. For example, in Python, you would use the ord and chr functions to convert to and from an integer, respectively.

# Common errors

This is the easiest problem in the contest. However, that doesn't mean nothing can go wrong with your submission. Lots of things can go wrong, from major problems down to simple gaffes such as choosing the wrong language while submitting! You must be prepared to handle these mistakes, because debugging is part of a programmer's life, and in fact usually takes more time than actual coding.

One common error is forgetting to initialize variables, as in the following:

#include <stdio.h>

char s[1111];
int main() {
int cases, cas, i, total = 0;
scanf("%d", &cases);
for (cas = 0; cas < cases; cas++) {
scanf("%s", s);
for (i = 0; s[i]; i++) {
if ('0' <= s[i] && s[i] <= '9') total += s[i] - '0';
}
printf("%d\n", total);
}
return 0;
}


This one could be quite misleading because the line total = 0 initializing the total variable is present. It's just that it's outside the loop, so it doesn't reset to 0 for the following test case! Consequently, your program will most likely get wrong answers for all but the first test case.

One way to detect this is possibly to add an identical test case in your sample input. If the answers you got are different, then most likely something's wrong. Also, to be sure, declare variables only on scopes they're needed, as shown in the C++ and Java examples above.

Another error is allocating an array too small for the input string. Remember that when given constraints such as $|S| \le 1000$, judges will most likely test out extremes, so allocating an array of size 200 will not be enough. In fact, for C or C++ an array of size $1000$ might not be enough too, because strings in those languages are terminated with a null character \0. You need at least $1001$ for the null character to be accommodated. (In fact, the examples above use $1111$ just to be extra cautious.)

• When hardcoding the digits, don't write the ASCII values. Instead, use the actual character '0', 'a', etc. Technically, '0' would work the same in most cases as 48 ('0''s ASCII code), but '0' is much clearer than a random 48 found in code.
• Simplicity is important. There's no need for very advanced programming patterns for something as simple as this. If you check the examples above, you find that they're simple and straight to the point, and it also has the desirable consequence of being clear and readable.
• Some submissions convert chars to ints by first converting the char to a string and then parsing it into an int. Please don't do that. It's harder to read than the methods above, and is actually much more inefficient.
• If possible, try to use library functions that will help you with the task if you know them. In actual real life coding work, it always pays to check whether a particular task already has a builtin function for it or has already been implemented in a library. In our case, you could use Character.isDigit in Java or isdigit in C++ to check whether a char is a digit.

# Miscellanea

Here we mention other common errors and things to remember.

• Don't forget to print a newline at the end of each test case! To be sure, try it yourself first with a couple of test cases and see if they output as expected. Command line tools such as diff or fc are helpful for this task and guards against manual checking which is unreliable.
• Use zero-indexing. It's idiomatic in many languages including C, C++ and Java. Even though some problems use one-indexing, it's usually helpful to convert to zero indexing in most languages, especially if you're gonna use some builtin functions that assume zero indexing.
• Don't print unnecessary stuff like "Please enter the number of test cases: ". These are considered part of your output and so your solution will be marked wrong. The goal is to exactly match the contents of the judge's output file.

Language-specific things

• When using scanf to take a string, use scanf("%s",s); instead of the incorrect scanf("%s",&s);.
• When copy-pasting Java code to be submitted to CodeChef, use Main as your public class name (as stated in https://www.codechef.com/wiki/sample-solutions ). It's because Java requires the file name and public class name to be the same, but since you copy-pasted code, CodeChef doesn't know the public class name and could have a harder time figuring it out from code. It's easier to just assume it's something, say, Main.

Contest-specific stuff

These are things usually good to be done during contests but are worth keeping in mind that they are considered bad practice in industry code.

• In Java, there are times when some methods have checked exceptions, such as IOException for BufferedReader.readLine(), requiring you to handle them with bulky try-catch blocks. In these cases, you can just declare throws Exception in your main method. This is very much not recommended in production code, but in a contest, it's very convenient.
• In Java, instead of importing each class separately, you can import everything from a single package using "import *", e.g. import java.util.*;. Of course this is not recommended at work because it makes it harder to keep track of where each used class is defined, but during contests you don't really need to worry about it.
• Instead of having to type std::cin all the time, you can add using namespace std in your C++ code so you can just use cin. It's also considered bad practice in real life. Nonetheless, it's also convenient, makes code a bit clearer, and can cut down typing time. Even then, people using it must be wary, because problems such as this and this might occur.

# Time Complexity:

$O(N)$

# AUTHOR'S AND TESTER'S SOLUTIONS:

This question is marked "community wiki".

1.7k586142
accept rate: 11%

1

great editorial for beginners to learn how to write better code!!

(21 Apr '18, 16:11)
 toggle preview community wiki:
Preview

By Email:

Markdown Basics

• *italic* or _italic_
• **bold** or __bold__
• image?![alt text](/path/img.jpg "title")
• numbered list: 1. Foo 2. Bar
• to add a line break simply add two spaces to where you would like the new line to be.
• basic HTML tags are also supported
• mathemetical formulas in Latex between \$ symbol

Question tags:

×15,852
×1,191
×354
×59
×17

question asked: 01 Dec '15, 21:52

question was seen: 1,586 times

last updated: 21 Apr '18, 19:08