clickhouse-cli

An unofficial command-line client for the ClickHouse DBMS. It implements some common and awesome things, such as:

Autocompletion (work in progress)
Syntax highlighting for the queries & data output (Pretty* formats)
Multiquery & multiline modes by default - paste anything as much as you want!
Pager support (less) for the data output
Custom, PostgreSQL-like commands like \d+ table_name or \ps. See \?
User-defined functions

But it works over the HTTP port, so there are some limitations for now:

Doesn't fully support sessions. SET options are stored locally and are sent with every request.

Install

Python 3.4+ is required.

$ pip3 install clickhouse-cli

Options

$ clickhouse-cli --help
Usage: clickhouse-cli [OPTIONS] [SQLFILE]

  A third-party client for the ClickHouse DBMS.

Options:
  -h, --host TEXT          Server host (hostname, or URL)
  -p, --port INTEGER       Server HTTP port
  -u, --user TEXT          User
  -P, --password           Password
  -d, --database TEXT      Database
  -s, --settings TEXT      Query string to be appended to every query
  -c, --cookie TEXT        Cookie header to be sent with every query
  -f, --format TEXT        Data format for the interactive mode
  -F, --format-stdin TEXT  Data format for stdin/file queries
  -m, --multiline          Enable multiline shell
  --stacktrace             Print stacktraces received from the server.
  --version                Show the version and exit.
  --help                   Show this message and exit.

Configuration file

~/.clickhouse-cli.rc is here for your service!

[defaults]
# Default connection options that will be used if the relevant argument was omitted.

host = 127.0.0.1
port = 8123
db = default
user = default
password =
cookie =

# It's not secure to store the password here in plain text.


[main]
# Disable multiline mode by default
multiline = False

# Show SQL statements timing
timing = True

# Preferred data format for the interactive mode
format = PrettyCompact

# Preferred data format for the non-interactive mode (file/stdin)
format_stdin = TabSeparated

# Show the reformatted query after its execution
show_formatted_query = True

# Syntax highlighting
highlight = True

# Syntax highlight certain output in the interactive mode:
highlight_output = True

# Syntax highlighting in TrueColor (if supported, see https://gist.github.com/XVilka/8346728)
highlight_truecolor = True

# Pygments Highlight theme (check out https://help.farbox.com/pygments.html for available themes)
highlight_theme = default

# Show the output via pager (if defined)
pager = False


[settings]
# You can place the server-side settings here!

# max_memory_usage = 20000000000

Quickstart

$ clickhouse-cli
clickhouse-cli version: 0.1.6
Connecting to localhost:8123
Connected to ClickHouse server.

 :) help

clickhouse-cli's custom commands:
---------------------------------
USE     Change the current database.
SET     Set an option for the current CLI session.
QUIT    Exit clickhouse-cli.
HELP    Show this help message.

PostgreSQL-like custom commands:
--------------------------------
\l      Show databases.
\c      Change the current database.
\d, \dt Show tables in the current database.
\d+     Show table's schema.
\ps     Show current queries.
\kill   Kill query by its ID.

Query suffixes:
---------------
\g, \G  Use the Vertical format.
\p      Enable the pager.

 :) \l

┌─name───────┐
│ default    │
│ quickstart │
│ system     │
└────────────┘

Ok. 3 rows in set. Elapsed: 0.022 sec.

 :) USE quickstart

Changed the current database to quickstart.

Ok.

 :) \dt

┌─name───┐
│ ontime │
└────────┘

Ok. 1 row in set. Elapsed: 0.012 sec.

 :) SELECT OriginCityName, count(*) AS flights
    FROM ontime GROUP BY OriginCityName ORDER BY flights DESC LIMIT 5

┌─OriginCityName────────┬──flights─┐
│ Chicago, IL           │ 10536203 │
│ Atlanta, GA           │  8867847 │
│ Dallas/Fort Worth, TX │  7601863 │
│ Houston, TX           │  5714988 │
│ Los Angeles, CA       │  5575119 │
└───────────────────────┴──────────┘

Ok. 5 rows in set. Elapsed: 1.317 sec.

Advanced usage

Environment variables

The available environment variables are:

CLICKHOUSE_HOST
CLICKHOUSE_PORT
CLICKHOUSE_USER
CLICKHOUSE_PASSWORD
CLICKHOUSE_DATABASE
CLICKHOUSE_COOKIE

The order of precedence is:

command argument
environment variable
default value in the ~/.clickhouse-cli.rc

Reading from file / stdin

$ echo 'SELECT 1, 2, 3; SELECT 4, 5, 6;' | clickhouse-cli
1	2	3

4	5	6

$ cat test.sql
SELECT 1, 2, 3;
SELECT 4, 5, 6;

$ clickhouse-cli test.sql
1 2 3

4 5 6

$ clickhouse-cli -F CSV <<< 'SELECT 1, 2, 3 UNION ALL SELECT 4, 5, 6'
1,2,3
4,5,6

Inserting the data from file

$ clickhouse-cli -q 'CREATE TABLE test (date Date, s String, i UInt64) ENGINE = TinyLog'

$ cat data.csv
2017-01-01,hello,1
2017-02-02,world,2

$ clickhouse-cli -q 'INSERT INTO test (date, s, i)' -F CSV data.csv

Ok. Elapsed: 0.037 sec.

$ clickhouse-cli -q 'SELECT * FROM test'
2017-01-01	hello	1
2017-02-02	world	2

Custom settings

$ clickhouse-cli -h 10.1.1.14 -s 'max_memory_usage=20000000000&enable_http_compression=1'

User-defined functions

Oh boy. It's a very dirty (and very untested) hack that lets you define your own functions or, actually, whatever you want, by running a find & replace operation over the query before sending the query to the server.

Say, you often run queries that parse some JSON, so you use visitParamExtractString all the time:

 :) SELECT date, ip, visitParamExtractString(headers, 'User-Agent') AS ua FROM visits LIMIT 1;

Even with autocompletion, this makes it harder to work with such queries. With this feature, you'll be able to create custom find & replace pairs to make things a little bit easier (or harder; it depends). Put this in your .clickhouse-cli.rc:

udf = {
        r'header\((.*?)\)': r'visitParamExtractString(headers, \1)',
  }

And rejoice!

 :) SELECT date, ip, header('User-Agent') AS ua FROM visits LIMIT 1;

The client will replace the matching expressions with another ones, and the query will execute correctly. See .clickhouse-cli.rc for a full example.

tpphu / clickhouse-cli