Generate nice ID base on MongoDB ObjectID about hashids HOT 5 CLOSED

I'm going to guess that you're trying to figure out a better way to compress Mongo's IDs because the current output is kinda lengthy.

There are two problems with your code that I see:

1. When you do preg_replace with '', you cannot safely encode/decode "5ab3" and "a53b". And when you do str_replace with range, you cannot safely decode because you don't know which integer was a letter (5a => 50 or 50 => 50?).
2. Additionally, your chunk_split will have an issue when you split a string like this: "200003", because encode will think you're encoding 200 and 3, when you should be encoding 200 and 003.

So, to answer your question -- yes, looks like you'll definitely have collisions with these methods.

from hashids.

55a3f9bdd874595813000029 => 920000318595478ddb9f3a55 => 920000318595478331953055

[0] => 920000318
[1] => 595478331
[2] => 953055
[3] => time(1)

Right now i;m not care about decrypt ID back, need be only uniq, so i think to add mt_rand ord time(1) as additional param. Also MongoDB ids are normal ended with 0000001 so i just strrev base ID. So som kind of hard-fix.

Thank you for your patience

from hashids.

If you need them to be unique, I don't recommend doing what you're proposing, because you will generate duplicate ids (irregardless of what additional parameters you add). I'm not aware of any safer method other than the one provided by encode_hex and decode_hex.

from hashids.

$id = new \MongoId($toHash);

echo $h->encode([$id->getTimestamp(),$id->getInc(),$id->getPID(),preg_replace('/[a-zA-Z]+/', '', substr($toHash,8,6))]);

55c1f1f4e12f92d814000031
Timestamp: 1438773748
Hostname2: 1292 //MD5 3 bytes :: -A_Z
ProcesID: 5336
Increment: 49
0DWpDj6ulYUL3ZuR5

Tested on 2milion rows no duplicate, so this method is i think more clean

from hashids.

Hi @Dawjan, yes that's one of the methods listed under http://hashids.org/#alternatives (number 2). That's good that you didn't get duplicates for the first 2m rows.

from hashids.