Where Python's regular expression engine is very robust, MySQL supports the two following metacharacters for forming regular expressions:

Pattern matching is always a matter of comparison. Therefore, with either of these, never use operators of equality.

SELECT * FROM menu WHERE name = 's%';                   WRONG
SELECT * FROM menu WHERE name <> 's%';            WRONG

Instead, use the keywords LIKE and NOT LIKE.

SELECT * FROM menu WHERE name LIKE 's%';                RIGHT
SELECT * FROM menu WHERE name NOT LIKE 's%';    RIGHT

Using metacharacters, one can match records using very irregular terms. Some of the possible combinations follow below:

For a smaller dataset or even larger datasets served over low-contest or no-contest connections (for example local servers or dedicated LAN connections), there is the option of running a live query to present the user with the possible options. If the user has specified the database and table to be used, as in the example seen previously, then it is a small matter to match patterns in a column using LIKE and a regular expression.

The MySQL sentence for what we are doing, along with its results, is as follows:

mysql> SELECT name FROM menu WHERE name LIKE 's%';
+--------+
| name |
+--------+
| salmon | 
| sole | 
+--------+
2 rows in set (0.00 sec)

In Python, the preceding statement would look like this:

column = 'name'
term = 's%'
statement = """select %s from menu where name like '%s'""" %(column, term)

Using the conversion specifier (%s), this code can easily be adapted for more dynamic uses.

Having restricted the parameters of the search, we are in greater control of the results and can therefore anticipate the number of fields in each record returned. We then have to execute the query and tell the cursor to fetch all of the records. To process the records, we iterate over them using a pattern similar to what we used previously:

command = cur.execute(statement)
results = cur.fetchall()

column_list = []
for record in results:
    column_list.append(record[0])

print "Did you mean:"
for i in xrange(0, len(column_list)):
    print "%s.  %s" %(i+1, column_list[i])
option = raw_input ('Number:')
intoption = int(option)

The results for this code are:

Did you mean:
1. salmon
2. sole
Number:

Naturally, we must then test the user input. After that, we can process the query and return the results.

This example is shown using terminal options so we do not use any JavaScript to transfer the options. However, in modern day reality, any application that relies on a web browser—either for background processing or for a primary interface, can use this code with minor modifications.