Updated 2016-08-03 10:13:13 by Kroc

mysqltcl is a simple interface to MySQL for Tcl

Attributes  edit

current version
3.052
release time
2012-10

Binary packages  edit

For Windows
http://www.xdobry.de/mysqltcl/mysqltcl-3.03.zip
For Mac OSX
http://www.zolli.fr/fichiers/mysqltcl-3.03.zip

See Also  edit

adb, by Roalt Aalmoes
a simple tcl database engine that uses mysqltcl underneath.

Documentation  edit

An introduction to the TclMySQL library, Diego Alberto Arias Prad, 2004-03

Description  edit

Tcl interface to mysql relational database. Supports Tcl 8 objects, unicode, nested queries, etc. Currently at version 3.05.

Example 1  edit

SS 2005-02-12: I wrote this very short example of mysqltcl usage for a guy asking on the tclers chat about how to use it. I'll put it here to redirect to this page the next time people ask about it.

This code connects to the mysql server at 'localhost', logs in with the username 'root', password 'foobar', selects the database 'mysql', and executes the query "SELECT HOST FROM USER". Every host returned is printed using puts, and finally the mysql handle is closed.
package require mysqltcl
set m [mysqlconnect -user root -db mysql -password foobar]
mysqluse $m mysql
foreach res [mysqlsel $m {select host from user} -flatlist] {
    puts $res
}
mysqlclose $m

Artur Trzewik: Some comments about example:

mysqluse $m mysql is not needed in this case because the database is already specified as connection parameter.

The fastest solution in mysqltcl for fetching data (bigger volumes) is mysqlmap. mysqlsel with -flatlist will build a Tcl-List with the whole result. For very big data-volumes the best choice is mysql::receive, as that does not use client caching (on the level of mysqlclient C-library). Starting with version 3.00, you can also use namespace command names.
#using fetch 
mysql::sel $m {select host from user}
while {[llength [set row [mysql::fetch $m]]]>0} {
    # row is always a list. using lindex avoids conversion list to string 
    puts [lindex $row 0]
}

# the fastest way to operate data without building big internal Tcl lists.
mysql::sel $m {select host from user}
mysql::map $m host {
    puts $host
}

# for very big data volumes. No caching at all
mysql::recieve $m {select host from user} host {
    puts $host
}

LV 2007-09-10: So, is it really mysql::recieve rather than mysql::receive ?

Example 2, test transaction  edit

Str 20141119-2248-strobel: I wrote a bigger sample code for c.l.tcl testing the transactional abilites of the storage engines. Enjoy.
#!/usr/bin/env tclsh

package require mysqltcl

proc getdb {} {
   global thisisthedbhandle
   if {![info exists thisisthedbhandle]} {
      set thisisthedbhandle [::mysql::connect -db yourdb -user you -password yourpass]
      ::mysql::autocommit $thisisthedbhandle false
   }
   return $thisisthedbhandle
}
proc getflat {sql} {
   return [::mysql::sel [getdb] $sql -flatlist]
}
proc dbcmd {sql} {
   return [::mysql::exec [getdb] $sql]
}
proc create1 {eng} {
  dbcmd "create table if not exists t1 (id integer auto_increment primary key, 
                   val integer not null default 0) engine = $eng"
  dbcmd "create table if not exists t2 (counter integer) engine = $eng"
  dbcmd "insert into t2(counter) values (0);"
}
proc drop1 {} {
  dbcmd "drop table if exists t1;"
  dbcmd "drop table if exists t2;"
}
proc showt {t} {
  switch $t {
   t1 { foreach {i e} [getflat "select * from t1"] {
          puts "id: $i\tvalue: $e" }
       }
    t2 { foreach e [getflat "select * from t2"] { 
            puts "sum in table 2: $e"  }
       }
  }
}

proc testit {} {
  # indeed, there is no start transaction in the interface, just do it by hand
  dbcmd "start transaction;"
  dbcmd "insert into t1 (val) values (4);"
  dbcmd "update t2 set counter = counter + 4;"
  dbcmd "commit;"
  puts "checking values are there:"
  showt t1
  showt t2
  puts "insert into t1 then rollback - result:"
  dbcmd "start transaction;"
  dbcmd "insert into t1 (val) values (42);"
  # why should i do ::mysql::rollback, it is too much to write
  dbcmd "rollback"
  showt t1
  showt t2
}
#--------------------------------
puts "test with MyISAM"
create1 MyISAM
testit 
drop1

puts "test with InnoDB"
create1 InnoDB
testit 
drop1
# ------------- end of tests, and glorious evening..
proc sing {} {
  puts "
    You can't always get what you want,
    You can't always get what you want,
    But if you try sometime,
    You might find 
    You get what you need!
 "
}
sing

Mysqltcl and Starkits  edit

DrumBSD:

I found this code on a mailing list which wraps mysqltcl for Microsoft Windows into a starkit or starpack. You have to modify pkg_Index.tcl like this:
proc loadmysqltcl { dir } {
    set oldcwd [pwd]
    cd $dir
   
   foreach file [glob *.dll] {
       file copy -force $file c:/winnt/temp
   }
   cd c:/winnt/temp
   load libmysqltcl[info sharedlibextension]
   cd $oldcwd
}

package ifneeded mysqltcl 3.01 [list loadmysqltcl $dir]

Anyway, there's a problem when "c:/winnt/temp" doesn't exists. So I tried to put a $tcl_platform(osVersion) into pkgIndex.tcl but it doesn't work. Maybe when it loads mysqltcl, the tcl_platform array doesn't exist. Any tips to solve this problem?

JH: Did you make sure to declare tcl_platform global in the proc?

ramsan: I think that all Microsoft Windows have defined the variable TEMP. So:
proc loadmysqltcl { dir } {
    set oldcwd [pwd]
    cd $dir
   
   foreach file [glob *.dll] {
       file copy -force $file $::env(TEMP)
   }
   cd $::env(TEMP)
   load libmysqltcl[info sharedlibextension]
   cd $oldcwd
}
package ifneeded mysqltcl 3.01 [list loadmysqltcl $dir]

Misc  edit

[crouzilles]: How do you use placeholders in mysqltcl? Can you do {select * from mytable where name = ?}?