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
In IJ1 the ResultsTable API deprecates accessing arbitrary cell values by (col,row). All data access is meant to work by (columnName,row).
Currently IJ2 allows unnamed columns (or even two columns with the same name). We should change the table API to allow them to index via column name rather than column number. And we should put in checks so that setting column headers only allows unique column names (they could default to "1", "2", "3", etc.).
These changes would allow us to simplify the code in ResultsTableHarmonizer and also make it more correct.
The Table API moved to https://github.com/scijava/scijava-table. The API is stable and unlikely to change in the near future. There will be a scijava-table2 eventually which fixes various things such as using composition over inheritance w.r.t. the Java collections interfaces. But I am not convinced unique column header names is indeed something we want to enforce, even in the next incarnation of the API.
In IJ1 the ResultsTable API deprecates accessing arbitrary cell values by (col,row). All data access is meant to work by (columnName,row).
Currently IJ2 allows unnamed columns (or even two columns with the same name). We should change the table API to allow them to index via column name rather than column number. And we should put in checks so that setting column headers only allows unique column names (they could default to "1", "2", "3", etc.).
These changes would allow us to simplify the code in ResultsTableHarmonizer and also make it more correct.
Migrated-From: http://trac.imagej.net/ticket/1601
The text was updated successfully, but these errors were encountered: