DbSchemaCLI Universal Command Line Client

DbSchemaCLI is a univeral database command-line client, featuring:

  • Execute queries on multiple databases and save the result to file
  • Transfer data between multiple databases
  • Write Groovy scripts
  • Deliver data to end-users via email or FTP
  • Execute database tasks on schedule
  • Create database custom commands like 'disk usage' or 'locking' and run them on multiple databases simultaneously

DbSchemaCLI is part of the DbSchema installation kit.


Connect to the Database

Connect to the database using:

connection mysql -dbms MySql -h localhost -u dbschema -p dbschema12 -d sakila
download driver <rdbms>

Before creating database connections, you have to download the JDBC drivers for the databases we are working with. For this execute in DbSchemaCLI the command:

download driver <rdbms>
The 'rdbms' is one of database names from this list, case sensitive.

Example: download driver PostgreSQL or download driver MySql. The drivers will be downloaded in C:\Users\YourUser\.DbSchema\drivers. Alternatively, download the JDBC driver .jar files by yourself and place them in the /home/users/.DbSchema/drivers/<rdbms>

In the User Home Directory (C:\Users\YourUser or /home/YourUser ) DbSchemaCLI is creating a folder .DbSchema/cli. with a file init.sql. DbSchemaCLI is executing this script at startup. So is easy to define the database connections right in this file.

  • register driver <rdbms> <driverClass> <URL> <"default_parameters">
  • connection <name> <rdbms> <"parameters">


register driver PostgreSql org.postgresql.Driver jdbc:postgresql://<host>:<port>/<db> "port=5432"
connection pos1 PostgreSql "user=postgres password=secret host=localhost db=sakila"

register driver Oracle oracle.jdbc.OracleDriver jdbc:oracle:thin:@<host>:<port>:<db> "port=1521"
connection ora1 Oracle "user='sys as sysdba' password=secret host= db=prod"

The register driver command defines for each RDBMS (MySql,Postgres,Oracle...) the:

  • the Java driver class
  • the JDBC URL with tags <...>.
  • default values for same tags, like the port
This way of defining connections is similar to PostgreSQL configurations. In the advanced section, you will find examples of how to create a new connection from an existing one, by changing one connection parameter. We call this 'derive connection'.

The connection command defines the connection by setting the real values for each URL parameter. The RDBMS from the register driver command should match the RDBMS from the connection command.

The complete list of known drivers is the same as for DbSchema.

Execute Queries on Multiple Databases

This operation assumes that the queried tables exists on each of the connected databases. The connections 'pos1' and 'pos2' should be already defined.

connect pos1,pos2
SELECT * FROM sakila.city;

Create connection groups.

connection group production pos1,pos2
connect production
SELECT * FROM sakila.city;
Spool the result to a file.
connect production exclude my3
spool /tmp/result.csv
SELECT * FROM sakila.city;
spool off

Use Variables

Define DbSchemaCLI variables using the command 'vset'. Few variables are system variables. If the system variable replace_variables=true, all other variables will be replaced in the SQL queries. Use help to list all system variables. Execute show variables to list the defined variables.
vset cli.spool.separator=|
vset format.date=YYYY-MM-dd'T'hh:mm:ss'Z'
vset format.timestamp=YYYY-MM-dd'T'hh:mm:ss'Z'
vset cli.settings.replace_variables=true

vset myid=6230
connect db4
execute export.sql
Where the export.sql script is :
spool /tmp/export_&amp;myid.csv
select * from some_table where id=&amp;myid;

Transfer Data Between Databases

The next command transfers data from the specified database(s) into the currently connected database. The transfer command can use a query to read the data. The column names should match between the source and the target databases. There can be multiple source databases.
connect pos1
transfer into_table from pos2,pos3 using
select id, firstname, lastname from persons;

The transfer is running on a configurable number of threads.

vset cli.settings.transfer_threads=4

Execute 'help' to see all configurable parameters.

There is one reader thread and multiple writer threads. Each writer thread is writing a bunch of rows, then it commits the transaction. The writer threads are writing between 500 and 100.000 records, depending on the number of transferred columns ( 500 for more than 50 columns, 100.000 for 1 column ).

Write Groovy Scripts

Many database tasks require more logic than simple SQLs. We do this using Groovy, a pure Java scripting language with closures. Groovy code can be started with the keyword groovy and ends with the first line with only a slash '/'.

Everything which works in Java is working in Groovy, with these improvements:

  • Multiline Strings (started and end with triple quotes ): """...."""
  • ${...} expressions are evaluated
  • Closures sql.eachRow(...){r->...} are similar with Java Lambda
Here is an sample Groovy script:
connect db1

int days = 5
sql.eachRow( """
SELECT personid, firstname, lastname FROM persons p WHERE p.created &lt; sysdate-${days} AND
  EXISTS ( SELECT 1 FROM sport_club s WHERE s.personid=p.personid )
  ORDER BY personid ASC
""".toString() ) { r ->
println "${r.personid} ${r.firstname} ${r.lastname}"

Groovy scripts receive two objects :

  • 'sql' is the database connection as an SQL object.
  • 'connector' is the DbSchemaCLI object for a defined connection
Please check the Help / Code Samples in DbSchema for more Groovy sample scripts.

Chunk Database Updates and Deletes

Updating or deleting large amounts of data may require splitting the operation into smaller chunks. This because the large operation may cause disk or memory issues and may lead to database locks.

  • The 'update' example is using a nested loop with a Postgres.setResultSetHoldability.
  • The 'delete' example is loading id values into an ArrayList.
import java.sql.ResultSet
sql.setResultSetHoldability( ResultSet.HOLD_CURSORS_OVER_COMMIT)
def i = 0
sql.eachRow( "SELECT personid FROM persons".toString() ){ r ->
sql.execute("UPDATE persons SET ocupation='t' WHERE personid=?".toString(), [ r.personid ])
if ( i > 1000 ){
i = 0
print "."
Here is an deletion example:
List ids = new ArrayList()
sql.eachRow( 'SELECT id FROM large_table' ) { m ->
ids.add( m.id )
int cnt = 0;
for ( int id : ids ) {
sql.execute('DELETE FROM large_table WHERE id=?', [id] )
if ( cnt > 1000) {
  print '.'
  cnt = 0;

Create Custom Commands

Here we create a Postgres custom command to compute the disk usage. The command can be executed on multiple databases. Save this script into a file diskUsage.groovy, and use 'register command diskUsage.groovy as disk usage'.
def cli = new CliBuilder(usage:'disk usage ')
cli.h(longOpt:'help', 'Show usage information and quit')
def options = cli.parse(args)

if (options.'help') {

sql.execute("DROP TABLE IF EXISTS dbschema_disk_usage")
sql.execute("DROP TABLE IF EXISTS dbschema_disk_usage_logs")
sql.execute("CREATE TABLE dbschema_disk_usage( filesystem text, blocks bigint, used bigint, free bigint, percent text, mount_point text )")
sql.execute("CREATE TABLE dbschema_disk_usage_logs( blocks bigint, folder text )")

sql.execute("COPY dbschema_disk_usage FROM PROGRAM 'df -k | sed \"s/  */,/g\"' WITH ( FORMAT CSV, HEADER ) ")
sql.execute("UPDATE dbschema_disk_usage SET used=used/(1024*1024), free=free/(1024*1024), blocks=blocks/(1024*1024)")

String dbid = sql.firstRow( "SELECT current_database() as dbid").dbid
sql.execute("COPY dbschema_disk_usage_logs FROM PROGRAM 'du -m -s /data/${dbid}/pgsystem/pg_log | sed \"s/\\s\\s*/,/g\"' WITH ( FORMAT CSV ) ".toString())
sql.execute("COPY dbschema_disk_usage_logs FROM PROGRAM 'du -m -s /data/${dbid}/jobsystem/pg_log | sed \"s/\\s\\s*/,/g\"' WITH ( FORMAT CSV ) ".toString())

int logBlocks = sql.firstRow( "SELECT sum(blocks) as usage FROM dbschema_disk_usage_logs").usage

sql.eachRow( "SELECT * FROM dbschema_disk_usage"){ r->
if ( r.filesystem.startsWith( '/dev/mapper') ){
int freePercent = Math.round(r.free*100/(r.blocks))
String status = ( r.free < 30 && r.blocks > 1500 ) ? 'Very Low' : (( r.free < 50 && r.blocks > 1000) || r.free < 30 ) ? 'Low' : "Ok"
String logStatus = logBlocks > 1024 ? String.format( 'pg_logs %dM', logBlocks ) : ''
println String.format( '%10s   Free %3d%% ( %dG out of %dG ) %s', status, freePercent, r.free, r.blocks, logStatus )

sql.execute("DROP TABLE dbschema_disk_usage_logs")
sql.execute("DROP TABLE dbschema_disk_usage")

Create Connections Using Suffixes

There are cases when you have to manage large database clusters where you have primaries and standby databases. I used to manage a cluster where the primary database hostnames were like 'pos1.db', 'pos2.db', ... and standbys were 'pos1.sbydb', 'pos2.sbydb'...

For this I let this script to create the standby database connections:

import com.wisecoders.dbs.cli.connection.*

for ( int i=1;i <130; i++){
ConnectionFactory.createConnector( "pos${i}", "postgres", "user=postgres host=pos${i}.db db=sakila${i}" )

ConnectionFactory.addSuffix( new Suffix("sby"){
public Connector create( Connector con ){
String host = con.getHost();
if ( host.endsWith(".db") ) return duplicate( con, "host=" + host.replaceFirst(/\.db/, /.sbydb/))
return null
The script will receive a connection. If the connection is matching the name .db, will duplicate the Connector by creating a new parameters string, where .db is replaced with .sbydb.

Write Cronjob Scripts

DbSchemaCLI you can execute scheduled scripts regularly. If a script is failing, an email is sent to the configured emails.
vset cli.mail.server.host=internal.mailserver
vset cli.mail.server.user=admin
vset cli.mail.server.password=secret
vset cli.mail.from=Admin
vset cli.mail.alert=admin@databaseconsole.com
vset cli.cron.folder=/dbschemacli/cronscripts
On the machine where DbSchemaCLI is running, add in /etc/crontab this line, fixing fist the path to DbSchemaCLI
01 * * * * /usr/local/bin/dbschemacli -cron
And in /dbchemacli/cronscripts/ save the script files. Example:
DbSchemaCLI is checking the file names matching this pattern 'schedule_' followed by 'minute|hour|dayOfWeek|dayOfMonth|dayOfYear|month|year'. After this, a pair of brackets '()' containing one of the following is expected:
  • number (1)
  • a list of numbers (1,2,5)
  • a range (1-10)
  • an expression 'each<number>', for example: 'each5'.
If the scripts fails, an email is sent to the database admin. If a file has the '.groovy' extension, the script is executed using Groovy.

Upload Reports on [s]ftp Server

You can upload CSV output files on FTP or SFTP servers. The command is upload [-d] ftpURL fileName. -d states for removing the local file. We are using Apache Commons VFS with support for all protocols. Example:
upload sftp://ftpusser@ftp.ftphost.com/intoFolder report.csv

Store Passwords in a Separate File

For many reasons, you may need to store passwords outside the init.sql file, where the database connections are being defined. This would allow saving the init.sql file in GIT without making passwords public. Therefore is possible to store passwords in a separate file ~/.DbSchema/dbshell/pass.txt or ~/.pgpass (for compatibility with PostgreSQL standards). The file format is the same as the PostgreSQL pgpass

Here we set the password 'secret' for all hosts, port 1521 user 'postgres'.
The connection should specify password=<pgpass> in the connection parameters to make use of the password from the pass.txt or .pgpass file.

Floating License Server

Users who purchased a floating license, should install DbSchema on a separated machine, where to run the Floating License Server.

  • In command prompt or terminal start DbSchemaCLI
  • Load the floating license from file using register -f 'path-to-file'
  • Start the license server using license server -s
  • On the clients machines, using DbSchema, open Help / Registration Dialog / Floating License Server, and set the server host and port
  • In DbSchemaCLI you can view the active users using the commands license server -l and license server -i

DbSchemaCLI Start Parameters

DbSchemaCLI can be started with the following parameters:

  • '-cron' to start in crontab mode, details in the Cronjob Documentation.
  • A script path to execute the given script.