README updates

[ppazos]

Requires: sudo apt install libmysqlclient-dev
Before build.

After installing got:

$ gcc -o levenshtein.so -shared levenshtein.c mysql_config --include
/usr/bin/ld: /tmp/ccBcelTa.o: relocation R_X86_64_PC32 against symbol `levenshtein' can not be used when making a shared object; recompile with -fPIC
/usr/bin/ld: final link failed: Bad value
collect2: error: ld returned 1 exit status

Tried: $ gcc -o levenshtein.so -shared -fPIC levenshtein.c mysql_config --include

Worked OK, I can see the new .so file

[ppazos]

mysql> CREATE FUNCTION levenshtein RETURNS INT SONAME 'levenshtein.so';
ERROR 1126 (HY000): Can't open shared library 'levenshtein.so' (errno: 2 /usr/lib/mysql/plugin/levenshtein.so: cannot open shared object file: No such file or directory)

After build, there is the need to copy the .so file to the location mentioned on the error above.

$ sudo cp levenshtein.so /usr/lib/mysql/plugin/

mysql> CREATE FUNCTION levenshtein RETURNS INT SONAME 'levenshtein.so';
Query OK, 0 rows affected (0.00 sec)

[ppazos]

Just to confirm:

mysql>  select * from mysql.func;
+---------------------+-----+----------------+----------+
| name                | ret | dl             | type     |
+---------------------+-----+----------------+----------+
| levenshtein         |   2 | levenshtein.so | function |
| levenshtein_k       |   2 | levenshtein.so | function |
| levenshtein_k_ratio |   2 | levenshtein.so | function |
| levenshtein_ratio   |   2 | levenshtein.so | function |
+---------------------+-----+----------------+----------+

Testing: there is an error: levenshtein_ratio returns the same as levenshtein instead of 0..1

mysql> select levenshtein_ratio('MARIO','mario');
+------------------------------------+
| levenshtein_ratio('MARIO','mario') |
+------------------------------------+
|                                  5 |
+------------------------------------+
1 row in set (0.00 sec)

mysql> select levenshtein_ratio('Mario','mario');
+------------------------------------+
| levenshtein_ratio('Mario','mario') |
+------------------------------------+
|                                  1 |
+------------------------------------+
1 row in set (0.00 sec)

mysql> select levenshtein('Mario','mario');
+------------------------------+
| levenshtein('Mario','mario') |
+------------------------------+
|                            1 |
+------------------------------+
1 row in set (0.00 sec)

mysql> select levenshtein('MARIO','mario');
+------------------------------+
| levenshtein('MARIO','mario') |
+------------------------------+
|                            5 |
+------------------------------+
1 row in set (0.00 sec)

[ppazos]

In fact, all functions return the same as levenshtein, here tested lev_k_ratio:

mysql> select levenshtein_k_ratio('mario','Mario',1);
+----------------------------------------+
| levenshtein_k_ratio('mario','Mario',1) |
+----------------------------------------+
|                                      1 |
+----------------------------------------+
1 row in set (0.00 sec)

mysql> select levenshtein_k_ratio('mario','mArio',1);
+----------------------------------------+
| levenshtein_k_ratio('mario','mArio',1) |
+----------------------------------------+
|                                      1 |
+----------------------------------------+
1 row in set (0.00 sec)

mysql> select levenshtein('mario','mArio');
+------------------------------+
| levenshtein('mario','mArio') |
+------------------------------+
|                            1 |
+------------------------------+
1 row in set (0.00 sec)

[ppazos]

The only way to make the ratio work is by using levenshtein() /* the only function that works from mysql */ and calculating the ratio:

mysql> select (1- levenshtein('mario','Mario') / GREATEST(LENGTH('mario'), LENGTH('Mario')));
+--------------------------------------------------------------------------------+
| (1- levenshtein('mario','Mario') / GREATEST(LENGTH('mario'), LENGTH('Mario'))) |
+--------------------------------------------------------------------------------+
|                                                                         0.8000 |
+--------------------------------------------------------------------------------+
1 row in set (0.00 sec)