Printable View
i have 3 files which
contain data about people
in each file the lines are indexed by the social security number
so how find out that there are people which appear in the three files
i am not allowed to use rewind??
whats the algorithm for solving it?
Phew!! Maybe...are the files too big to just load them all into memory?Quote:
Originally Posted by transgalactic2
If all the files are in sorted order, you can read one record from each file. If they are not the same, then read more from whichever file has the lowest index number. If they are the same, do whatever you have to do when it's a 3-way match. Repeat until any file is empty.
--
Mats
if i will apply you tactics and go check a number of row each time
the first row in file 1 then first row in file 2 and first row in file three
and compare the data of each row
it will not wotk for this case:
005 appears in all file but by your method it will not show that 005 appeares in more then one file
??
file1:
000
001
002
003
005
file2:
003
004
005
file3:
005
006
007
first row check comparing 000 003 005
second row check comparing 001 004 006
etc..
so by your method it will show that 005 doesnt appear
in more then one file
Then you did not implement what I told you to implement. Show us what you've written.
If you read from the files you have:
file 1:
000
file 2:
003
file 3:
005
Next time you read file1, keeping the other values.
You would continue reading file1 until it gets to 003. It's still no match, so you read another from file1 (or file2 - doesn't really matter): 004. 003 from file2 (or file1) is now the lowest, so we read another from there too. We now have 004 and 004 and 005. Still no match. Read again from file1 and file2: 005 005 and 005 MATCH! Since the next read from file1 or file2 is EOF, we can stop.
--
Mats
ok i understand i need to pick the highest and read from the lower two indexes till you get an equal or bigger in which case
we need to pick the next biggest and pick from it
etc..
so in my exampe
in the first check 005 is the biggest the others are 003 and 000 i check if there is equal
i stay with 005 on file 3 and proggress the others 004 and 001 i check if there is equal
i stay with 005 on file 3 and proggress the others 005 and 002 i check if there is equal
but next i get EOF on file2 but there is 003 on file1
so we missed a match of 003 between file 1 and file 2
??
Hash the SSN, index a table of counters, each time you see an SSN bump the counter, at the end, any counter which == 3 corresponds to an SSN that was in all three files.
whats a SSN?
whats a HASH?
SSN is "Social Security Number". Hash is a mechanism to make a checksum/unique(ish) number from a text string or similar. I don't actually believe this is necessary.
As to the other post:
Matching between two files is the same as matching between three:
Read from the one that is lowest in the set. If you find the same number in two files, do whatever you are supposed to do, then go on and read new data from the file(s) with the lowest value.
--
Mats
i showed that by progressing the other two simultaneously
i miss matches.
so i need to progress the lowest each time??
Yes, you must only progress the lowest one.Quote:
Originally Posted by transgalactic2
--
Mats
i tried to find the smallest using these two operations
but still those "if" combination are too many
how to shorten the process of finding the smallest??
Code:if(strcmp(f1,f2)>0)
{
if(strcmp(f2,f3)>0)
{
//smallest is f3
}
}
Why not do it the way 974562348 (note: I made that number up) have done it in posted "minimum" codes on the internets? Find the smallest between one and two, and then check the answer to that against three.
but you see that i dont have such option as minimum in strcmp there is only
>0 or <0
and it depends on many cases
so again i have 6 if cases
how to shorten it?
after the pointed comment
i dont know from what file the smallest belongs to
i dont know in what two files i got the equality
so i could write it somewhere and go the the next line of the "smallest" file
??
Code:#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define false 0
#define true 1
char* equal(char*f1,char*f2,char*f3);
char* smallest(char*f1,char*f2,char*f3);
int main()
{
FILE *file1,*file2,*file3;
char f1[20];
char f2[20];
char f3[20];
char* small;
char* equall;
file1=fopen("c:\\1.txt","r");
file2=fopen("c:\\2.txt","r");
file3=fopen("c:\\3.txt","r");
fscanf(file1,"%9s",f1);
fscanf(file2,"%9s",f2);
fscanf(file3,"%9s",f3);
small=smallest(f1,f2,f3);
equall=equal(f1,f2,f3); //how to continue after this point
fclose(file1);
fclose(file2);
fclose(file3);
return 0;
}
char* smallest(char*f1,char*f2,char*f3)
{
char* pLowest = f1;
if(strcmp(pLowest, f2) > 0)
pLowest = f2;
if(strcmp(pLowest, f3) > 0)
pLowest = f3;
return pLowest;
}
char* equal(char*f1,char*f2,char*f3)
{
if(strcmp(f1,f3)==0)
{
return f1;
}
if(strcmp(f2,f3)==0)
{
return f2;
}
return NULL;
}
So, you need a function that actually tells you which file to read next, when you have three sets of SSN.
And of course, you need to do a loop of some sort.
--
Mats
how to return the file in this function
how to know what file is the smallest string
??
Code:
char* smallest_file(char*f1,char*f2,char*f3,FILE* file1,FILE* file2,FILE* file3)
{
char* pLowest = f1;
if(strcmp(pLowest, f2) > 0)
pLowest = f2;
if(strcmp(pLowest, f3) > 0)
pLowest = f3;
return pLowest;
}
is there any shorter way then
Code:char* smallest_file(char*f1,char*f2,char*f3,FILE* file1,FILE* file2,FILE* file3)
{
char* pLowest = f1;
if(strcmp(pLowest, f2) > 0)
pLowest = f2;
if(strcmp(pLowest, f3) > 0)
pLowest = f3;
if(strcmp(f1,pLowest)==0)
{
return file1;
}
if(strcmp(f2,pLowest)==0)
{
return file2;
}
if(strcmp(f3,pLowest)==0)
{
return file3;
}
}
Well, I suppose you could write a function that updates the lowest string(s) by reading from the relevant file, rather than making a function that returns the file wanted.
--
Mats
you say to write a function that updates the lowest string(s) by reading from the relevant file.
but i have to know from what file the lowest string came from.
so i cant understand how i can manage doing what you told me
without building such function?
If you keep the files and the strings in the same order at all times, you can determine if the value from file X needs updating, and read the file.
--
Mats
you say
build some how that you could know which file to update each time
but its too abstract
what sub operations i need to build?
Shouldn't need many sub-operations. Just string compare and fscanf() or something like that, and a few if-statements.
I just hacked one up in about 13 lines of code, with 5 if-statements (including finding the lowest one).
--
Mats
ok here is the comparing part
here is the scanf partCode:char* smallest(char*f1,char*f2,char*f3,FILE* file1,FILE* file2,FILE* file3)
{
char* pLowest = f1;
if(strcmp(pLowest, f2) > 0)
pLowest = f2;
if(strcmp(pLowest, f3) > 0)
pLowest = f3;
return pLowest;
}
for what reason i need to useCode:fscanf(file1,"%9s",f1);
fscanf(file2,"%9s",f2);
fscanf(file3,"%9s",f3);
"a few if-statements"
You need if-statments to determine if you need to read from the file or not (if you already have a value higher than the lowest, you should not read the file).
--
Mats
ok i tried to write it
but how to know which string is the smallest and which file it belongs to
without writing the previous function that i showed
??
Code:#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define false 0
#define true 1
char* equal(char*f1,char*f2,char*f3);
char* smallest(char*f1,char*f2,char*f3);
int main()
{
FILE *file1,*file2,*file3;
char f1[20];
char f2[20];
char f3[20];
char* small;
char* equall;
file1=fopen("c:\\1.txt","r");
file2=fopen("c:\\2.txt","r");
file3=fopen("c:\\3.txt","r");
while(fgetc(f1)!=EOF)||(fgetc(f2)!=EOF)||(fgetc(f1)!=EOF))
{
fscanf(file1,"%9s",f1);
fscanf(file2,"%9s",f2);
fscanf(file3,"%9s",f3);
small=smallest(f1,f2,f3);
equall=equal(f1,f2,f3);
}
fclose(file1);
fclose(file2);
fclose(file3);
return 0;
}
char* smallest(char*f1,char*f2,char*f3,FILE* file1,FILE* file2,FILE* file3)
{
char* pLowest = f1;
if(strcmp(pLowest, f2) > 0)
pLowest = f2;
if(strcmp(pLowest, f3) > 0)
pLowest = f3;
return pLowest;
}
char* equal(char*f1,char*f2,char*f3)
{
if(strcmp(f1,f3)==0)
{
return f1;
}
if(strcmp(f2,f3)==0)
{
return f2;
}
return NULL;
}
how to continue??
how to combine those parts
How about using a dynamically allocated array of structs whose elements are: SSN, current file number and a count. SSNs whose count equals 3 are present in all of the files. Just my 2c.
i am not allowed to use data structures
i am allowed to have only the strings of one ow from each file
i got a problem in your algorithm
its copying 005 many times
instead of once
each time we update the smallest it
puts in the file
005 003 005 >>puts 005 in the file
005 004 005 >>puts 005 in the file
005 005 005 >>puts 005 in the file
and my while condition doesnt work
its not stoping in that point why??
file1
000
001
002
003
005
file2
003
004
005
file3
005
006
007
Code:#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#define false 0
#define true 1
int equal(char*f1,char*f2,char*f3);
int smallest(char*f1,char*f2,char*f3);
int main()
{
FILE *file[3];
FILE *common;
char f[3][20];
int small;
int equall;
file[0]=fopen("c:\\1.txt","r");
file[1]=fopen("c:\\2.txt","r");
file[2]=fopen("c:\\3.txt","r");
common=fopen("c:\\common.txt","w");
fscanf(file[0],"%9s",f[0]);
fscanf(file[1],"%9s",f[1]);
fscanf(file[2],"%9s",f[2]);
while(fgetc(file[0])!=EOF||fgetc(file[1])!=EOF||fgetc(file[2])!=EOF)
{
small=smallest(f[0],f[1],f[2]);
equall=equal(f[0],f[1],f[2]);
if(equall!=-1)
fputs(f[equall],common);
fputc('\n',common);
fscanf(file[small],"%20s",f[small]);
}
fclose(file[1]);
fclose(file[2]);
fclose(file[0]);
return 0;
}
int smallest(char* f_a,char* f_b,char* f_c)
{
int i=0;
char* pLowest = f_a;
if(strcmp(pLowest, f_b) > 0)
{
i=1;
pLowest = f_b;
}
if(strcmp(pLowest, f_c) > 0)
{
pLowest = f_c;
i=2;
}
return i;
}
int equal(char* f1,char*f2,char*f3)
{
if(strcmp(f1,f3)==0)
{
return 0;
}
if(strcmp(f1,f2)==0)
{
return 1;
}
if(strcmp(f2,f3)==0)
{
return 2;
}
return -1;
}