We don't bite newbies here... much PerlMonks

### Re: Converting a number back to it's original string (that was hashed to generate that number)

by roboticus (Chancellor)
 on Jan 23, 2013 at 13:28 UTC ( #1014903=note: print w/replies, xml ) Need Help??

Nik:

How many different strings do you want to convert back and forth? For a 5 digit number, the maximum possible set of strings that you could map back and forth is under a million. It's a pretty straightforward result of the Pigeonhole Principle.

So if you want to do this, you're going to have to come up with rules to constrain your set of input strings to a set small enough to fit in the number of pigeonholes you have. *AND* you'll have to come up with a unique mapping between those strings and the numbers. (You, not us.)

The instant you have the possibility of mapping two different strings into the same number, you have lost information, and can no longer tell the two strings apart from the number. There's no way to reliably know which string to map the number to without some other source of information.

roboticus

Sorry for the troll food.

• Comment on Re: Converting a number back to it's original string (that was hashed to generate that number)

Replies are listed 'Best First'.
Re^2: Converting a number back to it's original string (that was hashed to generate that number)
by Nik on Jan 23, 2013 at 13:44 UTC
The number of the .html pages are < 100.
Please DON'T tell me to save both the pin <=> filepath and associate them (that can be done by SQL commands, i know)
I will not create any kind of primary/unique keys to the database.
I will not store the filepath into the database, just the number which indicates the filepath(html page).
Also no external table associating fielpaths and numbers.
i want this to be solved only by Python Code, not database oriented.

That is: I need to be able to map both ways, in a one to one relation, 5-digit-integer <=> string

int( hex ( string ) ) can encode a string to a number. Can this be decoded back? I gues that can also be decoded-converted back because its not losing any information. Its encoding, not compressing.

But it's the % modulo that breaks the forth/back association.

So, the question is:

HOW to map both ways, in a one to one relation, (5-digit-integer <=> string) without losing any information?

Nik:

What part of my post was unclear? If you're going to persist, then you, not us, are going to have to figure out the mapping, ensure the constraints on the string, and write the python code.

I'll speak no more on the topic.

...roboticus

When your only tool is a hammer, all problems look like your thumb.

blue_cowdawg's code
my @i = split(//,\$url); # put each letter in it's own bin my \$j=0; # Initailize our my \$k=1; # hashing increment values my @m=(); # workspace foreach my \$n(@i){ my \$q=ord(\$n); # ASCII for character \$k += \$j; # Increment our hash offset \$q += \$k; # add our "old" value \$j = \$k; # store that. push @m,\$q; # save the offsetted value } my \$hashval=0; #initialize our hash value # Generate that map { \$hashval = (\$hashval + \$_) % 100000} @m;
Is this the solution i seek to turn an 'absolute path' <=> '5-digit number' ?

HOW to map both ways, in a one to one relation, (5-digit-integer <=> string) without losing any information?

CREATE TABLE pin_to_url ( pin int PRIMARY KEY, url varchar(255) ); SELECT v.*, pu.url FROM visitors v JOIN pin_to_url pu ON pu.pin = v.pin ORDER BY whatever;

Just store the filename alongside the pin into the table and it'll work pretty well.

The only other way of mapping string to pin -- apart from the lookup table I've just typed out -- is enumerating over the possible filenames and hashing each in turn, comparing the result to the pin. There's no way around it with the constraints you stubbornly add.

Oh, by the way, with 100 files and 10k available (random-ish) identification numbers, you might hit the birthday paradox. About 60% probable, I think. You don't want a hashing function -- you want an integer sequence. Databases provide those.

Create A New User
Node Status?
node history
Node Type: note [id://1014903]
help
Chatterbox?
 [Corion]: Cool. Google makes it easier for me to wean myself off Google Keep. The app now insists on saving items online before displaying them in the offline view. [Corion]: I should maybe set up a "public" instance of my clone just to manage feature suggestions ;)

How do I use this? | Other CB clients
Other Users?
Others examining the Monastery: (12)
As of 2017-04-25 11:50 GMT
Sections?
Information?
Find Nodes?
Leftovers?
Voting Booth?
I'm a fool:

Results (450 votes). Check out past polls.