/dbx

Redis Module for maintaining hash by simple SQL (Support csv import/export)

Primary LanguageCMIT LicenseMIT

Redis Module for maintaining hash by simple SQL (Support csv import/export)

This module aims to provide simple DML to manipulate the hashes in REDIS for SQL users. It works as simple as you expected. It translates the input statement to a set of pure REDIS commands. It does not need nor generate any intermediate stuffs which occupied your storages. The target data is your hashes only. It also provides the CSV import and export function.

Usage

$ redis-cli
127.0.0.1:6379> hmset phonebook:0001 name "Peter Nelson"     tel "1-456-1246-3421" birth "2019-10-01" pos 3 gender "M"
127.0.0.1:6379> hmset phonebook:0002 name "Betty Joan"       tel "1-444-9999-1112" birth "2019-12-01" pos 1 gender "F"
127.0.0.1:6379> hmset phonebook:0003 name "Bloody Mary"      tel "1-666-1234-9812" birth "2018-01-31" pos 2 gender "F"
127.0.0.1:6379> hmset phonebook:0004 name "Mattias Swensson" tel "1-888-3333-1412" birth "2017-06-30" pos 4 gender "M"
127.0.0.1:6379> dbx select name,tel from phonebook where gender = "F" order by pos desc
1) 1) name
   2) "Bloody Mary"
   3) tel
   4) "1-666-1234-9812"
2) 1) name
   2) "Betty Joan"
   3) tel
   4) "1-444-9999-1112"

Getting started

Get the package and build the binary:

$ git clone https://github.com/cscan/dbx.git
$ cd dbx/src && make

This plugin library is written in pure C. A file dbx.so is built after successfully compiled.

Load the module in redis (3 ways)

  1. Load the module in CLI
127.0.0.1:6379> module load /path/to/dbx.so
  1. Start the server with loadmodule argument
$ redis-server --loadmodule /path/to/dbx.so
  1. Adding the following line in the file redis.conf and then restart the server
loadmodule /path/to/dbx.so

If you still have problem in loading the module, please visit: https://redis.io/topics/modules-intro

More Examples

Select statement

You may specify multiple fields separated by comma

127.0.0.1:6379> dbx select name, gender, birth from phonebook
1) 1) name
   2) "Betty Joan"
   3) gender
   4) "F"
   5) birth
   6) "2019-12-01"
2) 1) name
   2) "Mattias Swensson"
   3) gender
   4) "M"
   5) birth
   6) "2017-06-30"
3) 1) name
   2) "Peter Nelson"
   3) gender
   4) "M"
   5) birth
   6) "2019-10-01"
4) 1) name
   2) "Bloody Mary"
   3) gender
   4) "F"
   5) birth
   6) "2018-01-31"

"*" is support

127.0.0.1:6379> dbx select * from phonebook where birth > '2019-11-11'
1)  1) "name"
    2) "Betty Joan"
    3) "tel"
    4) "1-444-9999-1112"
    5) "birth"
    6) "2019-12-01"
    7) "pos"
    8) "1"
    9) "gender"
   10) "F"

If you want to show the exact keys, you may try rowid()

127.0.0.1:6379> dbx select rowid() from phonebook
1) 1) rowid()
   2) "phonebook:1588299191-764848276"
2) 1) rowid()
   2) "phonebook:1588299202-1052597574"
3) 1) rowid()
   2) "phonebook:1588298418-551514504"
4) 1) rowid()
   2) "phonebook:1588299196-2115347437"

The above is nearly like REDIS keys command

127.0.0.1:6379> keys phonebook*
1) "phonebook:1588298418-551514504"
2) "phonebook:1588299196-2115347437"
3) "phonebook:1588299202-1052597574"
4) "phonebook:1588299191-764848276"

Each record is exactly a hash, you could use raw REDIS commands hget, hmget or hgetall to retrieve the same content

Where clause

Your could specify =, >, <, >=, <=, <>, != or like conditions in where clause. Now the module only support "and" to join multiple conditions.

127.0.0.1:6379> dbx select tel from phonebook where name like Son
1) 1) tel
   2) "1-888-3333-1412"
2) 1) tel
   2) "1-456-1246-3421"
127.0.0.1:6379> dbx select tel from phonebook where name like Son and pos = 4
1) 1) tel
   2) "1-888-3333-1412"

Order clause

Ordering can be ascending or descending. All sortings are alpha-sort.

127.0.0.1:6379> dbx select name, pos from phonebook order by pos asc
1) 1) name
   2) "Betty Joan"
   3) pos
   4) "1"
2) 1) name
   2) "Bloody Mary"
   3) pos
   4) "2"
3) 1) name
   2) "Peter Nelson"
   3) pos
   4) "3"
4) 1) name
   2) "Mattias Swensson"
   3) pos
   4) "4"
127.0.0.1:6379> dbx select name from phonebook order by pos desc
1) 1) name
   2) "Mattias Swensson"
2) 1) name
   2) "Peter Nelson"
3) 1) name
   2) "Bloody Mary"
4) 1) name
   2) "Betty Joan"

Top clause

127.0.0.1:6379> dbx select top 3 name, tel from phonebook order by pos desc
1) 1) name
   2) "Mattias Swensson"
   3) tel
   4) "1-888-3333-1412"
2) 1) name
   2) "Peter Nelson"
   3) tel
   4) "1-456-1246-3421"
3) 1) name
   2) "Bloody Mary"
   3) tel
   4) "1-666-1234-9812"
127.0.0.1:6379> dbx select top 0 * from phonebook
(empty list or set)

Into clause for copy hash table

You could create another hash table by into clause.

127.0.0.1:6379> dbx select * into testbook from phonebook
1) testbook:1588325407-1751904058
2) testbook:1588325407-1751904059
3) testbook:1588325407-1751904060
4) testbook:1588325407-1751904061
127.0.0.1:6379> keys testbook*
1) "testbook:1588325407-1751904061"
2) "testbook:1588325407-1751904059"
3) "testbook:1588325407-1751904058"
4) "testbook:1588325407-1751904060"
127.0.0.1:6379> dbx select * from testbook
1)  1) "name"
    2) "Mattias Swensson"
    3) "tel"
    4) "1-888-3333-1412"
    5) "birth"
    6) "2017-06-30"
    7) "pos"
    8) "4"
    9) "gender"
   10) "M"
2)  1) "name"
    2) "Peter Nelson"
    3) "tel"
    4) "1-456-1246-3421"
    5) "birth"
    6) "2019-10-01"
    7) "pos"
    8) "3"
    9) "gender"
   10) "M"
3)  1) "name"
    2) "Bloody Mary"
    3) "tel"
    4) "1-666-1234-9812"
    5) "birth"
    6) "2018-01-31"
    7) "pos"
    8) "2"
    9) "gender"
   10) "F"
4)  1) "name"
    2) "Betty Joan"
    3) "tel"
    4) "1-444-9999-1112"
    5) "birth"
    6) "2019-12-01"
    7) "pos"
    8) "1"
    9) "gender"
   10) "F"

Into csv clause for exporting records in csv format

127.0.0.1:6379> dbx select * into csv "/tmp/testbook.csv" from phonebook where pos > 2
1) Kevin Louis,111-2123-1233,2009-12-31,6,F
2) Kenneth Cheng,123-12134-123,2000-12-31,5,M
127.0.0.1:6379> quit
$ cat /tmp/testbook.csv
Kevin Louis,111-2123-1233,2009-12-31,6,F
Kenneth Cheng,123-12134-123,2000-12-31,5,M
$

Delete statement

You may also use Insert and Delete statement to operate the hash. If you does not provide the where clause, it will delete all the records of the specified key prefix. (i.e. phonebook)

127.0.0.1:6379> dbx delete from phonebook where gender = F
(integer) 2
127.0.0.1:6379> dbx delete from phonebook
(integer) 2

Insert statement

The module provides simple Insert statement which same as the function of the REDIS command hmset. It will append a random string to your provided key (i.e. phonebook). If operation is successful, it will return the key name.

127.0.0.1:6379> dbx insert into phonebook (name,tel,birth,pos,gender) values ('Peter Nelson'     ,1-456-1246-3421, 2019-10-01, 3, M)
"phonebook:1588298418-551514504"
127.0.0.1:6379> dbx insert into phonebook (name,tel,birth,pos,gender) values ('Betty Joan'       ,1-444-9999-1112, 2019-12-01, 1, F)
"phonebook:1588299191-764848276"
127.0.0.1:6379> dbx insert into phonebook (name,tel,birth,pos,gender) values ('Bloody Mary'      ,1-666-1234-9812, 2018-01-31, 2, F)
"phonebook:1588299196-2115347437"
127.0.0.1:6379> dbx insert into phonebook (name,tel,birth,pos,gender) values ('Mattias Swensson' ,1-888-3333-1412, 2017-06-30, 4, M)
"phonebook:1588299202-1052597574"
127.0.0.1:6379> hgetall phonebook:1588298418-551514504
 1) "name"
 2) "Peter Nelson"
 3) "tel"
 4) "1-456-1246-3421"
 5) "birth"
 6) "2019-10-01"
 7) "pos"
 8) "3"
 9) "gender"
10) "M"
127.0.0.1:6379>

Note that Redis requires at least one space after the single and double quoted arguments, otherwise you will get Invalid argument(s) error. If you don't want to take care of this, you could quote the whole SQL statement by double quote as below:

127.0.0.1:6379> dbx "insert into phonebook (name,tel,birth,pos,gender) values ('Peter Nelson','1-456-1246-3421','2019-10-01',3, 'M')"

From clause for importing CSV file

The module provides simple import function by specifying from clause in Insert statement. It only support comma deliminated. Please make sure that the specified import file can be accessed by Redis server.

$ cat > /tmp/test.csv << EOF
"Kenneth Cheng","123-12134-123","2000-12-31","5","M"
"Kevin Louis","111-2123-1233","2009-12-31","6","F"
EOF
$ redis-cli
127.0.0.1:6379> dbx insert into phonebook (name, tel, birth, pos, gender) from "/tmp/test.csv"
1) "phonebook:1588509697-1579004777"
2) "phonebook:1588509697-1579004778"
127.0.0.1:6379> dbx select name from phonebook
1) 1) name
   2) "Kenneth Cheng"
2) 1) name
   2) "Kevin Louis"
127.0.0.1:6379> dbx delete from phonebook
(integer) 2
127.0.0.1:6379> quit
$ cat > /tmp/testheader.csv << EOF
name,tel,birth,pos,gender
"Kenneth Cheng","123-12134-123","2000-12-31","5","M"
"Kevin Louis","111-2123-1233","2009-12-31","6","F"
EOF
$ redis-cli
127.0.0.1:6379> dbx insert into phonebook from "/tmp/testheader.csv"
1) "phonebook:1588509697-1579004779"
2) "phonebook:1588509697-1579004780"
127.0.0.1:6379> dbx select name from phonebook
1) 1) name
   2) "Kenneth Cheng"
2) 1) name
   2) "Kevin Louis"

Issue command from BASH shell

$ redis-cli dbx select "*" from phonebook where gender = M order by pos desc
1)  1) "name"
    2) "Mattias Swensson"
    3) "tel"
    4) "1-888-3333-1412"
    5) "birth"
    6) "2017-06-30"
    7) "pos"
    8) "4"
    9) "gender"
   10) "M"
2)  1) "name"
    2) "Peter Nelson"
    3) "tel"
    4) "1-456-1246-3421"
    5) "birth"
    6) "2019-10-01"
    7) "pos"
    8) "3"
    9) "gender"
   10) "M"
$ redis-cli dbx select name from phonebook where tel like 9812
1) 1) name
   2) "Bloody Mary"

Note that "*" requires double quoted otherwise it will pass all the filename in current directory. Of course you could quote the whole SQL statement.

$ redis-cli dbx "select * from phonebook where gender = M order by pos desc"

Compatibility

REDIS v4.0

License

MIT

Status

This project is in an early stage of development. Any contribution is welcome :D