The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.


sqlt-graph - Automatically create a graph from a database schema


  ./sqlt-graph -d|--db|-f|--from=db_parser [options] schema.sql


    -l|--layout        Layout schema for GraphViz
                       ("dot," "neato," "twopi"; default "dot")
    -n|--node-shape    Shape of the nodes ("record," "plaintext,"
                       "ellipse," "circle," "egg," "triangle," "box,"
                       "diamond," "trapezium," "parallelogram," "house,"
                       "hexagon," "octagon," default "record")
    -o|--output        Output file name (default STDOUT)
    -t|--output-type   Output file type ("canon", "text," "ps," "hpgl,"
                       "pcl," "mif," "pic," "gd," "gd2," "gif," "jpeg,"
                       "png," "wbmp," "cmap," "ismap," "imap," "vrml,"
                       "vtx," "mp," "fig," "svg," "plain," default "png")
    -c|--color         Add colors
    --cluster          Cluster tables
    --no-fields        Don't show field names
    --height           Image height (in inches, default "11",
                       set to "0" to undefine)
    --width            Image width (in inches, default "8.5",
                       set to "0" to undefine)
    --fontsize         custom font size for node and edge labels
    --fontname         name of custom font (or full path to font file) for
                       node, edge, and graph labels
    --nodeattr         attribute name and value (in key=val syntax) for
                       nodes; this option may be repeated to specify
                       multiple node attributes
    --edgeattr         same as --nodeattr, but for edge attributes
    --graphattr        same as --nodeattr, but for graph attributes
    --natural-join     Perform natural joins
    --natural-join-pk  Perform natural joins from primary keys only
    --show-datatypes   Show datatype of each field
    --show-sizes       Show column sizes for VARCHAR and CHAR fields
    --show-constraints Show list of constraints for each field
    -s|--skip          Fields to skip in natural joins
    --skip-tables      Comma-separated list of table names to exclude
    --skip-tables-like Comma-separated list of regexen to exclude tables
    --debug            Print debugging information
    --trace            Print parser trace info


This script will create a graph of your schema. Only the database driver argument (for SQL::Translator) is required. If no output file name is given, then image will be printed to STDOUT, so you should redirect the output into a file.

The default action is to assume the presence of foreign key relationships defined via "REFERNCES" or "FOREIGN KEY" constraints on the tables. If you are parsing the schema of a file that does not have these, you will find the natural join options helpful. With natural joins, like-named fields will be considered foreign keys. This can prove too permissive, however, as you probably don't want a field called "name" to be considered a foreign key, so you could include it in the "skip" option, and all fields called "name" will be excluded from natural joins. A more efficient method, however, might be to simply deduce the foreign keys from primary keys to other fields named the same in other tables. Use the "natural-join-pk" option to achieve this.

If the schema defines foreign keys, then the graph produced will be directed showing the direction of the relationship. If the foreign keys are intuited via natural joins, the graph will be undirected.

Clustering of tables allows you to group and box tables according to function or domain or whatever criteria you choose. The syntax for clustering tables is:



Ken Youens-Clark <>.


perl, SQL::Translator.