You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is a feature request for various ways to combine Table objects. Specifically, it would be useful to have the option to:
append two tables (similar to append in ATpy, but more robust to slight differences in types). This would be useful for two tables that have the same columns, and where one table should be appended after the other.
Those tables could be even more flexible in merging than that. They could be merged by closest distance in coordinates (once a table column holds the coordinates as in #664) or even more complicated functions e.g. "find all sources with in 1 arcsec and match the new row to the brightest of those sources, even if that is not the closest match"
To that end, it would be useful, if the join could be done with an arbitrary function, that, given the new row and the table data, returns the matching indices or, in a different implementation, with an arbitrary function that, given the new row and the table data, returns a "cost" or "chi^2" for each table row.
@hamogu - yes, but if you are interested in the other kind of joining (e.g. source matching) that you mentioned in your previous comment, feel free to open a new issue.
Activity
hamogu commentedon Jan 21, 2013
Those tables could be even more flexible in merging than that. They could be merged by closest distance in coordinates (once a table column holds the coordinates as in #664) or even more complicated functions e.g. "find all sources with in 1 arcsec and match the new row to the brightest of those sources, even if that is not the closest match"
To that end, it would be useful, if the join could be done with an arbitrary function, that, given the new row and the table data, returns the matching indices or, in a different implementation, with an arbitrary function that, given the new row and the table data, returns a "cost" or "chi^2" for each table row.
embray commentedon Jan 23, 2013
Custom merge strategies would be a great idea, though will probably require some finesse.
hamogu commentedon Jun 17, 2013
@astrofrog #903 provided the join, vstack and hstack methods.
Should this be closed?
astrofrog commentedon Jun 18, 2013
@hamogu - yes, but if you are interested in the other kind of joining (e.g. source matching) that you mentioned in your previous comment, feel free to open a new issue.
Merge pull request astropy#268 from apooravc/enumerate-pkgs