Convert UTF-8 string to alphanumeric string without information loss

Question

I want to use Jake Wharton's DiskLruCache for Android to cache CouchDb documents on disk. CouchDb ids are just any JSON String, so could look Sömething/Like/Thís. However, the library's docs state

Each cache entry has a string key and a fixed number of values. Each key must match the regex [a-z0-9_-]{1,64}.

So I need a way to transform an arbitrary strings to conform to the regex [a-z0-9_-]{1,64}, while still being unique. How can I do this elegantly?

Should the result still be (at least somewhat) readable? I mean, assuming mainly ASCII characters where used, should that be converted? — Thomas Mueller, Jan 21 '14 at 19:42
No, and I may also note: I do not have to be able to convert it back. So a hash should be ok. — Maarten, Jan 21 '14 at 19:47

score 1 · Accepted Answer · answered Jan 21 '14 at 19:43

1

How about calculating a 64 character hash of the original JSON String and using this hash as a key for the cache?

But, this would not be guaranteed to be unique. But then again, mapping any JSON String to *[a-z0-9_-]{1,64}* will never be anyways.

answered Jan 21 '14 at 19:43

Tobias

4,999
7
34
40

True, true! So I would I generate this hash in Java? – Maarten Jan 21 '14 at 19:48
1

Yes, take a look at http://stackoverflow.com/questions/2624192/good-hash-function-for-strings for string hashing examples. – Tobias Jan 21 '14 at 19:56

score 0 · Answer 2 · edited May 23 '17 at 11:49

0

From this question: you can convert the original string to a string representation of the hexidecimal representation of its bytes.

public String toHex(String arg) {
    return String.format("%040x", new BigInteger(1, arg.getBytes("UTF-8")));
}

Although this might produce a String significantly larger than the previous, and may overflow 64-character limit.

edited May 23 '17 at 11:49

Community

1
1

answered Jan 21 '14 at 19:46

Maarten

6,894
7
55
90

Jaffa · Answer 3 · 2014-01-21T19:45:29.873

-1

Using a hash is not feasible if you want to reverse it. Base 64 doesn't match your requirements, but you can try something similar:

Encode each character using only [a-z0-9_-], or more exactly, if the character doesn't match [a-z0-9_], replace it by its unicode value preceeded by -.

edited Jan 21 '14 at 19:45

answered Jan 21 '14 at 19:37

Jaffa

12,442
4
49
101

is it a-z or a-zA-Z ? If you can use caps also, then just replace + and / by _ and - – Jaffa Jan 21 '14 at 19:41

Convert UTF-8 string to alphanumeric string without information loss

3 Answers3