7.1.9 1d_err, 2d_err, ...: Cartesian Matching with Per-Object Errors

matcher=2d_err values*='<x> <y> <error>'
               params='<scale>'
               tuning='<bin-factor>'

values*:

• x: Cartesian co-ordinate #1
• y: Cartesian co-ordinate #2
• error: Per-object error radius

params:

• scale: Rough average of per-object error distance; just used for tuning in conjunction with bin factor

tuning:

• bin-factor: Scaling factor to adjust bin size; larger values mean larger bins

The 1d_err, 2d_err, ... matchers compare positions in N-dimensional Cartesian space like the 1d, 2d matchers described in Section 7.1.6, except that the match radius can be different for each row. Rows are considered to match when the separation reckoned by Pythagoras between the x, y, ... positions is no larger than the sum of the two per-row error values. Matching in any number of Cartesian dimensions can be done by extending this syntax in the obvious way.

A variant form 1d_err_q, 2d_err_q, ... does the same thing but combines the two per-row error values in quadrature rather than by summation. In that case the separation beween the two positions must be no larger than sqrt(error1²+error2²).

The scale parameter must be supplied, and should be approximately the characteristic size of the per-object error values. In conjunction with the bin-factor tuning parameter its value affects the performance of the match, but not the result.