Introduction

Redis client binding for gawk, the GNU implementation of awk.

It's features includes

• no global variable pollusion
• convinient functions with variable-length parameter list, though limited
• multiple redis connection supported

Warning!!! Since awk/gawk was not designed to handle binary data, it will be difficult to handle binary data in Redis server. I haven not tested thorougly.

Usage

Currently, redis.awk provides following functions:

• redis_open(HOST, PORT): to connect to the Redis server,
• redis_close(CONN): to close the connection to the Redis server,
• redis_command(CONN, RESP, ARGC, ...): to request Redis operation,
• redis_dump(RESP): (for debugging purpose) to dump the response from the Redis server.

Note that you should provide the number of arguments, ARGC to redis_command(). For example, if you want to request SET KEY VAL to the Redis server, you should provide 3 as ARGC parameter like:

redis_command(conn, resp, 3, "SET", "KEY", "VAL")

Look at the example1 file for the demonstration.

Pipeline

Currently, I'm experimental several implementaion choices.

Look at the example file especially, example2-3. Basically, you'll canned the Redis commands using redis_append(), then call redis_flush() to send to the Redis server, then, call redis_resp() multiple times exactly the same numbers that you called redis_append().

 redis_append(conn, reqs, 2, "GET", "name")
 redis_append(conn, reqs, 2, "INCR", "count")
 ...
 redis_flush(conn, reqs)
 
 for (i = 0; i < reqs["nreq"]; i++) {
   redis_resp(conn, resp)    # Get response Nth command
   # process 'resp' if needed
 }

redis_append() may send the Redis command to the server even if you didn't call redis_flush() if the queued command exceeds the internal limit. You may control the internal limit, by modifying REDIS["pipesize"]. Before modifying it, please read read the comment in the function definition of redis_append().