tdb - a text database

tdb is an in-memory non-relational "Text DataBase" (tdb) and reporting language that uses N-dimensional AVL trees. to approximate a sparse, N-dimensional hyper-cube.

Here are some trivial examples to illustrate a few of the most basic tdb features and capabilities.

In particular, note how the foreach and sort statements are used. Full details can be found in the tdb Users' Guide.

Download tdb from github.

tdb Usage Message

Expenditure Fields File (exp.f)

Expenditure Data File (exp.d)

Expenditure by Category and Penguin - a Trivial Report Program

Running tdb in report mode:

This report output consists of rows of expenditure by category with a column for each penguin. It would be much more useful if it had a title, column headers, totals by category, totals by penguin and if the categories were further broken down by expenditure description. For example...

Expenditure by Category, Description and Penguin

Note the use of sum, count and number to perform the calculations - without using any variables! Also, the categories are sorted by the total category spend (ascending), the descriptions within each category are sorted by those totals (ascending), and the penguins are sorted by their total spend (ascending).

Expenditure by Category and Date

In this report, sort is used to order the sum of expenditure by category, decreasing.

Note the totals are sorted descending. This report output has a column for every date. Obviously, it would be too wide for normal data (with lots of dates). To address this issue, tdb in select mode can be used to select only the data of interest.

Expenditure Select File (exp_1_week.s)

This is the same report as above, but this time tdb is called twice - first in select mode (to choose only the required data - 1 week's worth in this example) and that data is piped to a second tdb instance running in report mode:

Again, the category expenditure total column is sorted in descending order. Also note that in the report above (with all the data), "Train Ticket" was before "Coffee" - but they have swapped position in this report (because the category totals are different).

Expenditure by Category, Description and Date

This report sorts by the sum of the expenditure based on the descriptions field, ascending - within each category.

Expenditure by Penguin, Category and Date (active dates)

By default (when sort is not used), every field's data is sorted ascending in alphabetical (for string fields) or numerical (for int and float fields) order.

Note that the date columns (in the output above) are not the same for all penguins. While this format might be useful in some cases, the report output might be more readable if all the dates were displayed for every penguin. This would require a trivial modification to the above program (see below).

Expenditure by Penguin, Category and Date - Report Source (all dates)

Now all penguins have the same dates. Some "cells" are empty, but this output is more readable.

Some Other Very Useful Features - Grocery Store Prices

Grocery Fields File (gro.f)

Comments in the fields file.

Grocery Data File (gro.d)

Note that there can also be comments in the data files, blank lines are ignored, and the same item has occasionally been purchased from the same "store".

The new field "store" is is created by concatenating the chain and location fields (see below).

Also, the data need not be aligned (as it is in these example data files); it must be separated by "white space".

Grocery Report Source

Using the join statement to effectively create new fields from the original ones can be particularly useful - as in this example.

This report also makes use of some other features that are (in this report) egregiously unnecessary. In particular, the use of the #include and multi-dimensional arrays could very easily have been avoided.

Grocery Report Include File (gro.i)

Grocery Report Output

Please refer to the tdb Users' Guide for more info.

Based entirely on the particularly speculative and overwhelmingly circumstantial evidence above, Tux may have thrown an outraaaaaaaaaaaageous party in Hong Kong on 6 Dec! (But, as we know, what happens on the road stays on the road.)