In some simplistic DBMS named PreQueL, the only column type allowed is CHAR(1) (a single character), and furthermore, its values are restricted to English upper-case letters ('A' to 'Z'). Table may contain up to 9 columns, numbered from 1 to 9. Tables themselves are named with lower-case English letters ('a' to 'z').

The only database query possible first joins all the tables, then selects some rows according to conditions in one of two forms: either < column>=< value> or < column1>=< column2>, for example a2=A or b1=c4. All conditions must hold simultaneously, as if they were connected by 'AND' operator.

You must write a PreQueL processor, which, given a tables and a set of conditions, will produce query result, i.e. those rows of a join satisfying all the conditions. Resulting rows must be sorted alphabetically.

The first line of input file contains of two integers - number of tables T and number of conditions D.

Starting from the second line there are T tables represented with number of rows RN and number of columns CN in the first line of a table, which is followed by RN lines consisting of exactly CN characters each. D lines with conditions follow the whole set of tables.

The constraints are: 1 <= T <= 26, 1 <= D <= 50, 1 <= CN <= 9, 1 <= RN <= 1000.

Output file contains result rows, one row per line. No input query will produce more than 1000 rows. The result rows are to be sorted alphabetically.

2 2
3 2
AX
BX
BY
2 3
ACD
BCC
a1=b1
a2=X

AXACD
BXBCC

Northeastern Europe 2000, Far-Eastern Subregion