ejc-sql

ejc-sql turns Emacs into a simple SQL client; it uses a JDBC connection to databases via clojure/java.jdbc lib.

You can use multiple connections at the same time. Autocompletion and basic formatting of SQL scripts are also available.

Installation
Configuration
Usage
List of keybindings & functions
Yasnippet
Troubleshooting
Requirements
License

Installation

To run Clojure, install Leiningen (assuming you have already installed Java 7+).
Add MELPA (if not yet present) to your package-archives list.

Then you can install ejc-sql with the following command:

M-x package-install [RET] ejc-sql [RET]

Configuration

Here is an full-fledged real-world ejc-sql configuration example: ejc-sql-conf.

First, load ejc-sql package:

(require 'ejc-sql)

Set httpd port

To achieve async SQL queries evaluation, both Emacs and JVM side is an HTTP client and HTTP server. Emacs as HTTP client via CIDER pass a SQL query to JVM and don't expect any data response from the database. The JVM part prints the result dataset to the file (in pain text table format). Then JVM as HTTP client notifies Emacs: "data printed into filepath, please refresh the output buffer". The JVM side port can be configured by related CIDER customizations, whereas the default Emacs side HTTP server port can be customized by clomacs-httpd-default-port variable (8080 by default):

(setq clomacs-httpd-default-port 8090) ; Use a port other than 8080.

Autocomplete

Install auto-complete e.g. by the following command: M-x package-install [RET] auto-complete [RET] Enable autocomplete for ejc-sql minor mode:

(require 'ejc-autocomplete)
(add-hook 'ejc-sql-minor-mode-hook
          (lambda ()
            (auto-complete-mode t)
            (ejc-ac-setup)))

Autocompletion is available for the following databases:

Oracle
MS SQL Server
PostgreSQL
MySQL
Informix
H2
SQLite

Autocompletion data is stored in the database structure cache. This cache is located on Clojure side, so it's global: the same database structure information is shared between different buffers connected to the same database. An attempt to autocomplete requires data from cache or lanches a thread aimed to create it. If Clojure side has the database structure cache, autocompletion variants are returned immediately. If not, the database structure cache creation process starts. It's async, so the process of Emacs is not blocked, and the user can move the point (cursor), edit SQL, and so on. If the user waits for autocompletion and doesn't move point (cursor) during this process, he will get autocompletion variants. In order to checkout the current database connection cache run ejc-print-cache.

Any successfully executed DDL query (CREATE, ALTER, DROP, RENAME) clears current connection cache, so next autocompletion attempt will recreate it. To clean the current connection cache manually, you can run ejc-invalidate-cache.

Fuzzy matching

Non-nil ejc-use-flx enables flx fuzzy matching engine for autocompletion. flx-ido is required in this case, it can be installed by your favorite approach. E.g. by MEPLA: M-x package-install [RET] flx-ido [RET]

(setq ejc-use-flx t)

To customize the minimum number of typed chars use flx for autocompletion, 2 by default:

(setq ejc-flx-threshold 2)

Company mode

Install company-mode e.g. by the following command: M-x package-install [RET] company [RET] Enable company-mode completion frontend for ejc-sql minor mode:

(require 'ejc-company)
(push 'ejc-company-backend company-backends)
(add-hook 'ejc-sql-minor-mode-hook
          (lambda ()
            (company-mode t)))

If you want to automatically start completion after inserting a dot for company-mode despite company-minimum-prefix-length is bigger than 0, set ejc-complete-on-dot to t:

(setq ejc-complete-on-dot t)

To show documentation quickhelp install company-quickhelp by: M-x package-install [RET] company-quickhelp [RET]

To activate company-quickhelp add the following to your .emacs:

(company-quickhelp-mode)

Minibuffer completion

By default standard completing-read is used as minibuffer the completion system. This is allow you to use it with any configured completion mechanism for example, vertico, ivy or helm. You can change this to ido-completing-read (ido) by editing ejc-completion-system and selecting ido:

(setq ejc-completion-system 'ido)

ElDoc

Enable ElDoc for ejc-sql minor mode:

(add-hook 'ejc-sql-minor-mode-hook
          (lambda ()
            (ejc-eldoc-setup)))

ElDoc for functions and procedures is available for the following databases:

Oracle
PostgreSQL
MySQL

Performance & output customization

ejc-set-fetch-size sets limit for the number of records to output (50 by default). Set to nil if you want to disable this limit.

ejc-set-max-rows sets the limit for the number of records to contain in ResultSet (99 by default). Set to nil if you want to disable this limit, or you can set it the same value as ejc-set-fetch-size to increase select query execution performance.

Any time your ResultSet is bigger than ejc-set-fetch-size you will receive messages like "Too many rows. Only 50 from 99+ are shown.". To inhibit this messages you can set ejc-set-show-too-many-rows-message to nil (t by default).

ejc-set-column-width-limit sets limit for outputing the number of chars per column (30 by default). The rest will be replaced by .... Set to nil if you want to disable this limit. This setting is applied to the text representation of any field type, but it is especially useful for varchar and CLOB fields.

ejc-set-use-unicode sets using unicode for grid borders, e.g. use ─┼─ instead of -+- (nil by default).

All these functions change Clojure variables, so if you want to change defaults, to avoid Clojure nREPL autolaunch on Emacs start, you should add them to the ejc-sql-connected-hook in your .emacs, e.g.:

(add-hook 'ejc-sql-connected-hook
          (lambda ()
            (ejc-set-fetch-size 50)
            (ejc-set-max-rows 50)
            (ejc-set-show-too-many-rows-message t)
            (ejc-set-column-width-limit 25)
            (ejc-set-use-unicode t)))

Current result set table minor-mode is orgtbl-mode. This mode provides some functionality for post-processing and browsing the query results.

(setq ejc-result-table-impl 'orgtbl-mode)

Alternatively, you can use a simple and bare result set mode to maximize the buffer performance by setting ejc-result-table-impl to 'ejc-result-mode.

If you want to see the full text of some field (e.g. the full text of CLOB field) despite ejc-set-column-width-limit, and your ejc-result-table-impl is 'ejc-result-mode you can select a single-record result set (e.g. SELECT * FROM table WHERE id = 1).

If you want to see the full text of some field with newlines in case of multiline fields, you should select single-record and single-column result set (e.g. SELECT field FROM table WHERE id = 1). So, you will get a field value as-is despite ejc-set-column-width-limit and ejc-result-table-impl.

To illustrate the description above here are some output examples of query results that depend on configuration.

Assume you have the following database (this example uses MySQL):

CREATE TABLE product (
  id    INT UNSIGNED  NOT NULL AUTO_INCREMENT PRIMARY KEY,
  name  VARCHAR(30)   NOT NULL,
  quantity INT,
  price DECIMAL(7,2),
  description VARCHAR(255)
);
INSERT INTO product (name, price, quantity, description)
VALUES ('socks', 1.25, 10, CONCAT('A sock is an item of clothing worn\n',
                                  'on the feet and often covering the\n',
                                  'ankle or some part of the calf.\n',
                                  'Some type of shoe or boot is\n',
                                  'typically worn over socks.'));
INSERT INTO product (name, price, quantity, description)
VALUES ('sweater', 14.56, 5, CONCAT('A sweater, also called a jumper\n'
                                    'in British English, is a piece\n'
                                    'of clothing, typically with long\n'
                                    'sleeves, made of knitted or\n'
                                    'crocheted material that covers\n'
                                    'the upper part of the body.'));

output examples for orgtbl-mode (by default):

SELECT * FROM product

| id | name    | quantity | price | description                    |
|----+---------+----------+-------+--------------------------------|
|  1 | socks   |       10 |  1.25 | A sock is an item of clothi... |
|  2 | sweater |        5 | 14.56 | A sweater, also called a ju... |

SELECT * FROM product WHERE id = 1

| id | name  | quantity | price | description                    |
|----+-------+----------+-------+--------------------------------|
|  1 | socks |       10 |  1.25 | A sock is an item of clothi... |

SELECT description FROM product WHERE id = 1

| description                        |
|------------------------------------|
| A sock is an item of clothing worn |
| on the feet and often covering the |
| ankle or some part of the calf.    |
| Some type of shoe or boot is       |
| typically worn over socks.         |

output examples for ejc-result-mode:

SELECT * FROM product

id | name    | quantity | price | description
---+---------+----------+-------+-------------------------------
1  | socks   | 10       | 1.25  | A sock is an item of clothi...
2  | sweater | 5        | 14.56 | A sweater, also called a ju...

SELECT * FROM product WHERE id = 1

id          | 1
name        | socks
quantity    | 10
price       | 1.25
description | A sock is an item of clothing worn on the feet and often covering the ankle or some part of the calf. Some type of shoe or boot is typically worn over socks.

SELECT description FROM product WHERE id = 1

description
----------------------------------
A sock is an item of clothing worn
on the feet and often covering the
ankle or some part of the calf.
Some type of shoe or boot is
typically worn over socks.

Create connections interactively

The easiest way to create connections configuration is to use interactive connections creation.

In any sql-mode buffer run (C-c ei):

M-x ejc-connect-interactive <RET>

Then follow the creation steps: type your connection name, choose database type, host (or file path depends on selected database type), port, user name and password.

ejc-sql uses Aether API of Maven-resolver to automatically resolve and download the required JDBC driver (if not yet) for selected database type.

You can customize artifacts and their versions used as JDBC drivers for each database type in Leiningen format in ejc-jdbc-drivers custom variable.

After you type all required data a and new connection will be created, it will attempt to immediately connect current-buffer to this connection. Then you can use this connection name to connect from different buffers. Type (C-c ec):

M-x ejc-connect <RET> your-connection-name <RET>

This connection will be available during the current Emacs session. To keep it between Emacs restarts, you can open your .emacs file or any file, loaded as Emacs configuration, locate point (cursor) somewhere after (require 'ejc-sql) expression and run:

M-x ejc-insert-connection-data <RET> your-connection-name <RET>

This function inserts ejc-create-connection expression the same as you can accomplish via manual connection creation.

Create connections manualy

Install JDBC drivers

In most cases, you don't need to install JDBC drivers manually. Simply put, you can set a parameter :dependencies in ejc-create-connection function as a vector of the required artifacts in Leiningen format. In this case, ejc-sql will resolve, download (if not yet) all required jar dependencies via Aether and load them to CLASSPATH during the ejc-connect function run. E.g.:

(ejc-create-connection
 "Informix-db-connection"
 :dependencies [[com.ibm.informix/jdbc "4.50.3"]]
 ...
 )

Alternatively, you can pass the exact JDBC driver jar file in the :classpath parameter of ejc-create-connection function. E.g.:

(ejc-create-connection
 "Informix-db-connection"
 :classpath (concat "~/.m2/repository/com/ibm/informix/jdbc/4.50.3/"
                    "jdbc-4.50.3.jar")
 ...
 )

ejc-sql will try to resolve all required dependencies if this JBDC driver requires some dependencies itself anyway. But you can pass all requred dependencies manually as a vector of jar files paths. E.g.:

(ejc-create-connection
 "Informix-db-connection"
 :classpath (vector
              (concat "~/.m2/repository/org/mongodb/bson/3.8.0/"
                      "bson-3.8.0.jar")
              (concat "~/.m2/repository/com/ibm/informix/jdbc/4.50.3/"
                      "jdbc-4.50.3.jar"))
 ...
 )

The rest of this section describes how to download and install JDBC drivers manually. If you are familiar with JDBC, please omit it.

The most common way is to install JDBC drivers to your ~/.m2 directory. Here is a list of such installation examples. Anyway, it will become outdated soon, so please consult Google to install your database JDBC driver.

First of all, install Maven, then you can install your JDBC driver with one of the following commands.

Oracle

Download JDBC driver manually from oracle.com

Fix your actual JDBC version number -Dversion, filepath -Dfile and run command like this:

mvn install:install-file -Dfile="~/downloads/ojdbc7.jar" -DgroupId=com.oracle.jdbc -DartifactId=ojdbc7 -Dversion=12.1.0.2 -Dpackaging=jar -DgeneratePom=true

MS SQL Server

Download JDBC driver manually from microsoft.com

Fix your actual JDBC version number -Dversion, filepath -Dfile and run command like this:

mvn install:install-file -Dfile="~/downloads/sqljdbc.jar" -DgroupId=com.microsoft.sqlserver -DartifactId=sqljdbc -Dversion=6.0 -Dpackaging=jar -DgeneratePom=true

or from Maven Central:

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=com.microsoft.sqlserver:mssql-jdbc:6.2.2.jre8

JTDS

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=net.sourceforge.jtds:jtds:1.3.1

PostgreSQL

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=org.postgresql:postgresql:42.6.0

MySQL

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=mysql:mysql-connector-java:5.1.6

MariaDB

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=org.mariadb.jdbc:mariadb-java-client:1.1.7

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=com.h2database:h2:1.4.192

SQLite

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=org.xerial:sqlite-jdbc:3.8.11.2

Informix

mvn org.apache.maven.plugins:maven-dependency-plugin:get -Dartifact=com.ibm.informix:jdbc:4.50.3

Setup connections with ejc-create-connection function in your .emacs. It's first arg is your custom database connection name, the remaining args are the same as database connection structure of clojure/java.jdbc lib.

The configuration of ejs-sql might looks like this:

;; Create your JDBC database connections configuration:

MySQL connection

;; MySQL example
(ejc-create-connection
 "MySQL-db-connection"
 :classpath (concat "~/.m2/repository/mysql/mysql-connector-java/5.1.6/"
                     "mysql-connector-java-5.1.6.jar")
 :subprotocol "mysql"
 :subname "//localhost:3306/my_db_name"
 :user "a_user"
 :password "secret")

If you want to see MySQL-specific keywords in autocompletion list, please provide access to mysql.help_keyword table for your user, e.g.:

GRANT SELECT ON mysql.help_keyword TO a_user;

MariaDB connection

;; MariaDB example
(ejc-create-connection
 "MariaDB-db-connection"
 :dependencies [[org.mariadb.jdbc/mariadb-java-client "2.6.0"]]
 :classname "org.mariadb.jdbc.Driver"
 :connection-uri "jdbc:mariadb://localhost:3306/db_name"
 :user "a_user"
 :password "secret")

MS SQL Server connection

;; MS SQL Server example
(ejc-create-connection
 "MS-SQL-db-connection"
 :classpath (concat "~/.m2/repository/com/microsoft"
                     "/sqlserver/sqljdbc/4.2/sqljdbc-4.2.jar")
 :subprotocol "sqlserver"
 :subname "//localhost:1433"
 :user "a_user"
 :password "secret"
 :database "my_db_name")

;; MS SQL Server example (via URI)
(ejc-create-connection
 "MS-SQL-db-connection-uri"
 :classpath (concat "~/.m2/repository/com/microsoft"
                     "/sqlserver/sqljdbc/4.2/sqljdbc-4.2.jar")
 :connection-uri (concat "jdbc:sqlserver://localhost\\\\instance:1433;"
                         "databaseName=my_db_name;"
                         "user=a_user;"
                         "password=secret;"))

;; MS SQL Server example (via JTDS)
(ejc-create-connection
 "MS-SQL-db-connection-JTDS"
 :classpath (concat "~/.m2/repository/net/sourceforge/jtds"
                     "/jtds/1.3.1/jtds-1.3.1.jar")
 :connection-uri (concat "jdbc:jtds:sqlserver://localhost:1433/dbname;"
                         "instance=instance;"
                         "user=a_user;"
                         "password=secret;"))

Oracle connection

;; Oracle example (via Service Name)
(ejc-create-connection
 "Oracle-db-connection-sname"
 :classpath (concat "~/.m2/repository/com/oracle/jdbc"
                    "/ojdbc8/12.2.0.1/ojdbc8-12.2.0.1.jar")
 :dbtype "oracle"
 :dbname "my_service_name"
 :host "localhost"
 :port "1521"
 :user "a_user"
 :password "secret"
 :separator "/")

;; Oracle example (via SID)
(ejc-create-connection
 "Oracle-db-connection-sid"
 :classpath (concat "~/.m2/repository/com/oracle/jdbc"
                     "/ojdbc7/12.1.0.2/ojdbc7-12.1.0.2.jar")
 :dbtype "oracle:sid"
 :dbname "my_sid_name"
 :host "localhost"
 :port "1521"
 :user "a_user"
 :password "secret"
 :separator "/")

;; Oracle example (via URI)
(ejc-create-connection
 "Oracle-db-connection-uri"
 :classpath (concat "~/.m2/repository/com/oracle/jdbc"
                     "/ojdbc7/12.1.0.2/ojdbc7-12.1.0.2.jar")
 :connection-uri "jdbc:oracle:thin:@localhist:1521:dbname"
 :user "a_user"
 :password "secret"
 :separator "/")

H2 connection

;; H2 example
(ejc-create-connection
 "H2-db-connection"
 :classpath (file-truename
             "~/.m2/repository/com/h2database/h2/1.4.191/h2-1.4.191.jar")
 :subprotocol "h2"
 :subname "file://~/projects/my_proj/db/database;AUTO_SERVER=TRUE"
 :user "a_user"
 :password "secret")

;; H2 remote example
;; run on remote server first:
;; java -jar ~/.m2/repository/com/h2database/h2/1.4.192/h2-1.4.192.jar -tcpAllowOthers
(ejc-create-connection
 "H2-remote-db-connection"
 :classpath "~/.m2/repository/com/h2database/h2/1.4.192/h2-1.4.192.jar"
 :connection-uri (concat "jdbc:h2:tcp://192.168.0.1:9092/~/db/database;ifexists=true;"
                         "user=a_user;"
                         "password=secret;"))

SQLite connection

;; SQLite example
(ejc-create-connection
 "SQLite-conn"
 :classpath (concat "~/.m2/repository/org/xerial/sqlite-jdbc/"
                    "3.23.1/sqlite-jdbc-3.23.1.jar")
 :subprotocol "sqlite"
 ;; Use absolute path, e.g.:
 ;;   "file:///home/user/projects/my_proj/db/sqdb.db"
 ;;   "/home/user/projects/my_proj/db/sqdb.db"
 ;;   "file:///C:/Projects/my_proj/db/sqdb.db"
 ;;   "C:/Projects/my_proj/db/sqdb.db"
 ;; or expand it by file-truename (not applicable for Windows paths):
 :subname (concat "file://"
                  (file-truename "~/projects/my_proj/db/sqdb.db")))

PostgreSQL connection

;; PostgreSQL example
(ejc-create-connection
 "PostgreSQL-db-connection"
 :classpath (concat "~/.m2/repository/org.postgresql/postgresql/42.6.0/"
                    "postgresql-42.6.0.jar")
 :subprotocol "postgresql"
 :subname "//localhost:5432/my_db_name"
 :user "a_user"
 :password "secret")

Informix connection

;; Informix example
(ejc-create-connection
 "Informix-db-connection"
 :dependencies [[com.ibm.informix/jdbc "4.50.3"]]
 :classname "com.informix.jdbc.IfxDriver"
 :connection-uri (concat
                  ;; In the case of IPv6, ::1 should be used
                  ;; as the host instead of localhost.
                  "jdbc:informix-sqli://localhost:8201/test:"
                  "INFORMIXSERVER=myserver;"
                  "user=a_user;"
                  "password=secret;"))

Presto connection

;; Presto example
(ejc-create-connection
 "Presto-db-connection"
 :subprotocol "presto"
 :dependencies [[com.facebook.presto/presto-jdbc "0.232"]]
 :connection-uri (concat
                  "jdbc:presto://localhost:1234/dbName/schemaName?"
                  "user=a_user"))

ClickHouse connection

;; ClickHouse example
(ejc-create-connection
  "ch@180"
  :dependencies [[com.clickhouse/clickhouse-jdbc "0.3.2"]]
  :dbtype "clickhouse"
  :classname "com.clickhouse.jdbc.ClickHouseDriver"
  :connection-uri (concat "jdbc:clickhouse://10.1.4.180:8123/" "testdb"))

ElasticSearch connection

;; ElasticSearch example
(ejc-create-connection
  "es@177"
  :dependencies [[org.elasticsearch.plugin/x-pack-sql-jdbc "7.9.1"]]
  :dbtype "elasticsearch"
  :classname "org.elasticsearch.xpack.sql.jdbc.EsDriver"
  :connection-uri (concat "jdbc:es://172.16.13.177:9200/"))

Snowflake connection

(ejc-create-connection
 "snowflake"
 :dependencies [[net.snowflake/snowflake-jdbc "3.13.27"]
                [net.java.dev.jna/jna "5.13.0"]]
 :connection-uri (concat "jdbc:snowflake://my-db.snowflakecomputing.com:443"
                         "?user="my-email@myjob.com"
                         "&warehouse=my-warehouse"
                         "&role=my-role"
                         "&db=my-db"
                         "&authenticator=externalbrowser"
                         "&JDBC_QUERY_RESULT_FORMAT=JSON"))

Usage

Basic use case

First of all, open your SQL source file (or any sql-mode buffer).

On the other hand, there is a handy function to create temporary sql-mode buffers for playing with SQL: ejc-get-temp-editor-buffer. If you bind it, e.g. to:

(global-set-key (kbd "C-c eb") 'ejc-get-temp-editor-buffer)

then, when you press C-c eb, *ejc-sql-editor* buffer will be created; when you press M-1 C-c eb, *ejc-sql-editor-1* buffer will created and so on. This buffers can be saved as ordinary file buffers by save-buffer command to the appropriate files, located in ejc-temp-editor-file-path directory ("~/tmp/ejc-sql/" by default).

In any selected SQL buffer connect to your database:

M-x ejc-connect <RET> MySQL-db-connection <RET>

and wait until "Connected." message appears. This will add connection information to buffer local variables. Furthermore, if there is no ejc-sql dedicated Clojure REPL running, it will start it.

Since connection information is buffer-local, you should run ejc-connect for any new buffer. Any of ejc-sql-mode buffers can keep connection information to different databases and database types. But they use the same ejc-sql dedicated Clojure REPL to interact with databases via JDBC.

Then type your queries like this:

select something from my_table

and press C-c C-c to run it.

Have much fun!

Separators & delimiters

Use / char to separate expressions to evaluate (actually \n/), e.g.:

select something from my_table
/
select other from other_table

So, you don't need to select SQL snippet, simply put point (cursor) into code snippet and press C-c C-c (or desired keybinding). Borders of SQL will be found by Emacs buffer begin/end or this / separator.

It's possible to pass multiple statements, you can use ; delimiter to separate them:

insert into my_table (product, price) values ('socks', 1.25);
insert into my_table (product, price) values ('sweater', 14.56);
insert into my_table (product, price) values ('jeans', 25.30);
/
select * from my_table

Here, the first part is a single SQL snippet passed to ejc-sql backend but evaluated by 3 independent SQL statements (transactions). The output will looks like this:

Records affected: 1
Records affected: 1
Records affected: 1

Furthermore, you can change the delimiter inside SQL snippet. E.g. in this MySQL snippet ; replaced by $$ as transaction delimiter:

DELIMITER $$
CREATE PROCEDURE GetAllProducts()
BEGIN
  SELECT * FROM products;
END $$
/
CALL GetAllProducts();

Since ; symbols can be used very often as part of procedure syntax (e.g. in Oracle), you can disable splitting SQL code snippet to the sequence of separate transactions by setting :separator in DB connection configuration (see https://github.com/kostafey/ejc-sql#oracle-connection).

Use with org-mode

You can run M-x ejc-connect <RET> my-db-connection <RET> in org-mode buffers. In this case, major-mode will persists as org-mode, but all connection-related data will be added to the buffer.

* Create DB
** Product table
*** Create
#+begin_src sql
CREATE TABLE product (
  id    INT UNSIGNED  NOT NULL AUTO_INCREMENT,
  name  VARCHAR(30)   NOT NULL,
  price DECIMAL(7,2)
);
#+end_src

*** Fill
#+begin_src sql
INSERT INTO product (name, price) VALUES ('socks', 1.25);
INSERT INTO product (name, price) VALUES ('sweater', 14.56);
INSERT INTO product (name, price) VALUES ('jeans', 25.30);
#+end_src

*** Select
#+begin_src sql
SELECT * FROM product;
/
SELECT * FROM product WHERE name = 'jeans';
#+end_src

#+RESULTS:
: id | name    | price
: ---+---------+------
: 1  | socks   | 1.25
: 2  | sweater | 14.56
: 3  | jeans   | 25.30

Place point (cursor) into code snippet and run SQL statements via C-c C-c as always. For org-mode buffers code snippets borders considered as batch of SQL statement(s) boundaries. Furthermore, you can use ejc-sql-separator (/ by default) to divide batch of SQL statement(s) inside code block as in sql-mode buffers.

The SQL query evaluation result will be added to this org-mode buffer in #+RESULTS: section - an expected behaviour for org-mode users by default (see example above).

To avoid this behaviour and get results in popup window - as ejc-sql users expected, add to your .emacs:

(setq ejc-org-mode-show-results nil)

If your org-mode buffer connected via ejc-connect, any time you run C-c ' (org-edit-special) for code snippets, you will get new buffer with this minor-mode (ejc-sql-mode) and all connection-related data. So, you can operate inside it like in ordinary sql-mode buffer, which is already connected to the database.

You can use both ejc-sql and org-mode original org-babel execution engine simultaneously in one buffer.

To disable ejc-sql wrapper around org-mode SQL source code blocks, set ejc-org-mode-babel-wrapper to nil (enabled by default).

If ejc-org-mode-babel-wrapper is enabled and the current SQL source code block has a connection header arguments, you will be asked for confirmation.

Reference this discussion.

Use existing nREPL

Dedicated ejc-sql nREPL

If you have to restart Emacs multiple times, you can keep the ejc-sql Clojure backend alive between Emacs restarts by running this backend out of Emacs, and connect to it from Emacs.

To accomplish that, you should cd to your ejc-sql project folder (typically ~/.emacs.d/elpa/ejc-sql-<version>) and launch the nREPL via lein repl.

Then run in Emacs M-x ejc-connect-existing-repl, type Host and Port from your lein repl console output.

Finally, use M-x ejc-connect from any SQL buffer to connect to the exact database, as always.

Different project nREPL

You can use different nREPL for ejc-sql, e.g. if you develop a Clojure project via CIDER, you can use your project nREPL to interact with the database by JDBC and ejc-sql. To achieve this, enable using any CIDER nREPL for clomacs projects in your .emacs:

(setq clomacs-allow-other-repl t)

Then add ejc-sql to your project as a dependency in project.clj:

(defproject some-project "0.1.0-SNAPSHOT"
  ...
  :dependencies [[org.clojure/clojure "1.10.0"]
                 ...
                 [ejc-sql "0.4.1-SNAPSHOT"]]
  ...
  )

or if you don't want to change your project.clj file, you can add it globally in ~/.lein/profiles.clj, e.g.:

{:user {:plugins [[cider/cider-nrepl "0.25.0-alpha1"]]
        :dependencies [[ejc-sql "0.4.1-SNAPSHOT"]]}}

The actual version of ejc-sql backend in Clojars:

So, when you start your project nREPL via cider-jack-in, you can open any SQL file (sql-mode or org-mode buffer) and connect to the database by ejc-connect as usual and it will reuse the existing nREPL.

Goto definition & results ring

In terms of ejc-sql, SQL evaluation results can be result sets, record affected messages, SQL definition of entities or error messages.

Any SQL evaluation result saved to results ring - list of files ejc-sql-result-0.txt, ejc-sql-result-1.txt, and so on. They located in the TEMP folder, it can be customized by ejc-results-path. The number of files (number of previous results) can be customized by setting ejc-ring-length (10 by default).

You can see previous SQL evaluation result by C-M-b (ejc-show-prev-result) in *ejc-sql-output* buffer. To return back use C-M-f (ejc-show-next-result). This way you can navigate through the results ring.

Since *ejc-sql-output* buffer contains ejc-sql connection information, it makes possible to navigate through views & stored procedures code definitions. E.g. you can require ejc-describe-entity for some stored procedure, then require ejc-describe-entity inside *ejc-sql-output* for some stored procedure, used in this (just described) procedure definition, and so on. Then you can return to previous procedure definition by ejc-show-prev-result. So, it looks like goto definition, then return back. For the purpose of convenience, the following keybindings are provided:

M-. ejc-describe-entity
M-, ejc-show-prev-result

List of keybindings & functions

New keybindings defined in ejc-sql-mode minor mode:

Keyboard shortcut	Command	Description
`C-c e c`	`ejc-connect`	Select DB connection (configured by `ejc-create-connection`) and connect to it.
`C-c e i`	`ejc-connect-interactive`	Create new connection interactively and connect to it.
`C-c C-c`	`ejc-eval-user-sql-at-point`	Evaluate SQL script bounded by the `ejc-sql-separator` or/and buffer boundaries.
`C-c C-r`	`ejc-eval-user-sql-region`	Evaluate region selected SQL code.
`C-g`	`ejc-cancel-query`	Terminate current running query or run `keyboard-quit` if there is no running queries.
`C-h t`	`ejc-describe-table`	Describe SQL table.
`C-h d`	`ejc-describe-entity`	Get entity definition: show creation SQL of view, package, function, procedure or type.
`M-.`	`ejc-describe-entity`	Get entity definition: show creation SQL of view, package, function, procedure or type.
`M-,`	`ejc-show-prev-result`	Load previous SQL eval result in `ejc-sql-output` buffer.
`C-c e up`	`ejc-show-last-result`	Show last result.
`C-c e t`	`ejc-show-tables-list`	Show tables list.
`C-c e v`	`ejc-show-views-list`	Show views list.
`C-c e p`	`ejc-show-procedures-list`	Show stored procedures list.
`C-c e T`	`ejc-show-user-types-list`	Show user types list.
`C-c e s`	`ejc-strinp-sql-at-point`	Strip SQL (trim java string tokens).
`C-c e S`	`ejc-dress-sql-at-point`	Dress SQL (to copy-paste it to java code).
`C-c e f`	`ejc-format-sql-at-point`	Format (pretty-print) this SQL statement.
`C-M-b`	`ejc-previous-sql`	Goto previous SQL statement (or load previous SQL eval result in `ejc-sql-output`).
`C-M-f`	`ejc-next-sql`	Goto next SQL statement (or load next SQL eval result in `ejc-sql-output`).
`C-M-S-b`	`ejc-previous-sql`	Select from point to previous SQL statement.
`C-M-S-f`	`ejc-next-sql`	Select from point to next SQL statement.

List of other interactive functions

Command	Description
`ejc-connect`	Connect to database for current buffer
`ejc-quit-connection`	Close all database connections, quit Clojure REPL.
`ejc-format-sql-region`	Format (pretty-print) selected SQL snippet
`ejc-mark-this-sql`	Mark SQL script bounded by the `ejc-sql-separator` or/and buffer boundaries
`ejc-show-tables-list`	Show tables list
`ejc-show-constraints-list`	Show constraints list
`ejc-open-log`	Open log
`ejc-get-temp-editor-buffer`	Create ad-hoc SQL editor buffer, use prefix arg number to get many buffers
`ejc-print-cache`	Output current connection cache
`ejc-invalidate-cache`	Clean your current connection cache (database owners and tables list)
`ejc-direx:pop-to-buffer`	Create buffer with database structure tree

Yasnippet

List of snippets:

select	where	inner
insert	begin	left
update	grant	right
delete	revoke	alter

Troubleshooting

Error running timer ‘ac-update-greedy’: (error "Sync nREPL request timed out (op eval session...

Increase nrepl-sync-request-timeout, e.g.:

(setq nrepl-sync-request-timeout 60)

Requirements:

GNU Emacs 26.
Leiningen 2.x
clomacs
clojure/java.jdbc 0.5.8
dash
spinner.el
direx.el
auto-complete (optional)
company-mode (optional)
company-quickhelp (optional)
flx-ido (optional)
yasnippet (optional).

License

Distributed under the General Public License 2.0+

GabrielMMelo/ejc-sql