fixes a bug in forest() where study names are not shown

[ecortens]

the study_name column in the dataset is overridden and regenerated as the data is not passed to combine_data() from .combine_data_bi() which is called within forest() which causes the study names to be regenerated sequentially and therefore incorrectly shown on the plot; this seemed like the lowest impact way to fix it, just to pass the study names directly to combine_data() as an argument

[FBartos]

Hi @ecortens ,

Thanks for the report & fix. I am just in the mids of a large rewrite. The new version 4.0.0 will break backward compatibility—I won't be updating the 3.X version on CRAN anymore. But I will make sure the labels work in the new version.

Cheers,
Frantisek

[ecortens]

Hi @ecortens ,

Thanks for the report & fix. I am just in the mids of a large rewrite. The new version 4.0.0 will break backward compatibility—I won't be updating the 3.X version on CRAN anymore. But I will make sure the labels work in the new version.

Cheers, Frantisek

Wonderful!

Though my pull request doesn't speak to this, it would also be great if it were possible for forest() (or another function) to return the ORs that it calculates as part of the plotting process. Right now, to get them, the calculations are done in non exported functions, so you have to source() the source code and run them manually, which isn't ideal. Perhaps forest() could also simply return the data it plots, rather than just the plot.

[FBartos]

That would make sense. I'm also toying with an idea of creating a translation layer for the metafor's forest function so users could rely on that instead (and I would not need to maintain the forest plotting :).