Note:
tpipe
is experimental software. It is not tested
or documented as well as the rest of STIL, and may be
distributed within a separate package and/or radically modified
in future releases.
The current version is believed to work correctly, and I would
certainly like to know if you have problems, but essentially,
you are using it at your own risk!
Tpipe is a command-line tool for performing pipeline processing on tables. You can think of it as a supercharged tablecopy, or as a way of performing many of TOPCAT's table processing functions from the command-line, or as something like a Unix pipeline in which what gets passed from the input to the output is not a stream of bytes but the data and metadata of a table.
As described in the following sections, its capabilities include selecting rows using algebraic expressions based on column values as well as other criteria; adding new columns defined in terms of other columns using algebraic expressions, rearranging existing columns, sorting the rows, displaying table data and metadata, calculating column statistics, and of course reading and writing tables using any of STIL's supported table formats.
In most cases, processing is done in a stream-wise fashion, which means that arbitrarily large tables can be processed with minimal requirements of memory and usually quite fast. This is not always possible however; for instance a sort requires random access to the data values, so it might be necessary to read the whole input table before processing in that case.