Kathryn Andersen


sqlreport - make reports on a table in an SQLite database


version 0.16


sqlreport --help | --manpage | --version

sqlreport [ --all_pages ] --database database_file [ --distinct ] { --force_show colname=1 } { --groups template } { --headers template } { --header_start num } [ --index_template template ] [ --filename_format format ] [ --layout string ] [ --limit number ] [ --link_suffix string ] { --not_where colname=1 } [ --outfile filename ] [ --page number ] [ --report_style string ] { --row_ids table=colname } [ --report_template template ] [ --row_template template ] [ --prev_next_template template ] [ --multi_page_template template ] [ --split_ind_template template ] { --show colname } { --sort_by colname } { --sort_reversed colname=1 } [ --split_col colname [ --split_alpha number ] ] --table table [ --table_border number ] [ -- table_header string ] [ --title string ] [ --total ] [ --truncate_colnames number ] { --use_package pkgname } { --where colname=string }


This makes a report in HTML format, of a single table from an SQLite database. One can also create a non-HTML report if one gives a certain combination of options, but this is more oriented towards HTML reports.



Make a multi-page report, generating all pages, by page-number. The --limit and --outfile options are required for this.


The name of the database file to use. (required)


If columns are given to show (see show), then this will ensure that rows with exactly the same values will not be repeated.


An set of columns to always show in a row, even if they've already been shown in a header (see show).


Group template(s) (or filenames of files containing group templates). A group template is a template for values which are "grouped" under a corresponding header. The first group in the array is placed just after the first header in the report, and so on.

This argument can be repeated.

See headers for more information.


An array of header templates (or filenames of files containing header templates). A header template lays out what values should be put into headers rather than the body of the report. The first header template is given a H1 header, the second a H2 header, and so on. Headers are shown only when the value(s) they depend on change, but they get their values from each row in the report. Therefore the columns used in the headers should match the columns used in the sort_by array.

The column names are the variable names in this template. This has a different format to the report_template; it is more sophisticated.

The format is as follows:


A variable; will display the value of the column, or nothing if that value is empty.

{?colname stuff [$colname] more stuff}

A conditional. If the value of 'colname' is not empty, this will display "stuff value-of-column more stuff"; otherwise it displays nothing.

    {?col1 stuff [$col1] thing [$col2]}

This would use both the values of col1 and col2 if col1 is not empty.

{?colname stuff [$colname] more stuff!!other stuff}

A conditional with "else". If the value of 'colname' is not empty, this will display "stuff value-of-column more stuff"; otherwise it displays "other stuff".

This version can likewise use multiple columns in its display parts.

    {?col1 stuff [$col1] thing [$col2]!![$col3]}

The same format is used for groups and row_template and prev_next_template.


Print help message and exit.


Similar to the report_template, but this is used for the index-pages in multi-page and split reports. It has the same format, but it can be useful to have them as two separate templates as one may wish to change the way the title is treated for indexes versus actual reports.


The layout of the report. This determines both how rows are grouped, and what is in the generated row_template if no row_template is given.


The report is a (group of) tables, each row of the report is a row in the table; a new table occurs after the heading(s).


The report is in paragraphs, each row of the report is one paragraph.


The report is a (group of) lists, each row of the report is an item in the list; a new list occurs after the heading(s).


The rows are not HTML-formatted. The generated row_template is made up of Field:Value pairs, one on each line.


The rows are not HTML-formatted. The generated row_template is made up of values, one on each line.


The maximum number of rows to display per page. If this is zero, then all rows are displayed in one page.

The 'link_suffix' argument, if given, overrides the suffix given in links to the other pages in a multi-page report; this is useful if you're post-processing the files (and thus changing their extensions) or are using something like Apache MultiViews to eliminate the need for extensions in links.

    --link_suffix '.shtml'

    --link_suffix ''

Print the full help documentation (manual page) and exit.


A hash containing the column names where the selection criteria in where should be negated.


The name of the output file. If this is not given, or the name is '-' then the output goes to STDOUT.


Select which page to generate, if limit is not zero.


Template for previous and next links on multi-page reports.


The style of the report, especially as regards table layout.


Either a string containing a template, or string containing the name of a template file. The template variables are in the following format:


The following variables are set for the report:


Title (generally the table name).


The report itself.


The default column-name which identifies rows in SQLite is 'rowid', but for tables which have a primary integer key, this doesn't work (even though the documentation says it ought to). Therefore it is necessary to identify, for the given database, which tables need to use a different column-name for this. (This can be repeated)


The template for each row. This uses the same format as for headers. If none is given, then a default row_template will be generated, depending on what layout and which columns are going to be shown (see show).

Therefore it is important that if one provides a row_template, that it matches the current layout.

Also note that if a column is given in a header, it will not be displayed in a row, even if it is put into the row_template.


An array of columns to select; also the order in which they should be shown when a row_template has not been given. If this option is not used, all columns in the table will be shown.


An array of column names by which the result should be sorted. (Repeat the argument for each new value)


A hash of column names where the sorting given in sort_by should be reversed.


Generate a multi-page report where pages are split by the value of the given column (as well as by page-number if a limit is given)


If one is generating a split_col report, giving the 'split_alpha' option splits the report not by the distinct values of that column, but by truncated values of the column; giving a split_alpha value of 1 takes only the first letter, and so on.


The table to report on. (required)


For fine-tuning the report_style; if the layout is 'table', then this overrides the default border-size of the table.


When the report layout is 'table' and the report_style is not 'bare', then this argument can be used to customize the table-header of the report table. This must either contain the contents of the table-header, or the name of a file which contains the contents of the table-header.

If this argument is not given, the table-header will be constructed from the column names of the columns to be shown.


The title of the report; if this is empty, a title will be generated.


Just print the total matching rows, then exit.


For fine-tuning the report_style; this affects the length of column names given in layouts which use them, that is, 'table' (for all styles except 'bare') and 'para'. If the value is zero, the column names are not truncated at all; otherwise they are truncated to that number of characters.


Print informational messages.


Print version information and exit.


A hash containing selection criteria. The keys are the column names and the values are strings suitable for using in a GLOB condition; that is, '*' is a multi-character wildcard, and '?' is a single-character wildcard. All the conditions will be ANDed together.

Yes, this is limited and doesn't use the full power of SQL, but it's useful enough for most purposes.


An array of package names of packages to "use". This is mainly so that the {&funcname())} construct of the templates (see SQLite::Work::Template) can call functions within these packages (using their fully-qualified names).




perl(1) Getopt::Long Getopt::ArgvFile Pod::Usage


Please report any bugs or feature requests to the author.


    Kathryn Andersen (RUBYKAT)
    perlkat AT katspace dot com


Copyright (c) 2005 by Kathryn Andersen

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.