Issues with replace_na() in tidyr v1.2.0

[ArturoEsquerra]

Hi everyone!

We just noticed that a change in the newly released version of tidyr (v1.2.0) causes a problem with augsynth().

Specifically, in the new version of tidyr, the replace_na() function no longer allows the type of data to change (more info on this change can be found here). Therefore, replacing NA's of the integer column trt_time with Inf through the line: mutate(trt_time = replace_na(trt_time, Inf)) makes the code crash as replace_na() can't no longer convert from double to integer.

Best.
Arturo

[ebenmichael]

Thanks for catching this! Can you link to the relevant line? Would it fix things to make sure that the treatment time is a double?

[ArturoEsquerra]

Thanks for the quick reply! Matching the data types of trt_time and the replacement value should fix the issue.

I believe that the that the affected lines are:

1. Line 61 of augsynth_pre.R
2. Line 104 of format.R (potentially line 156 of the same .R file as well)

[idg23]

Is this issue solved? Thanks in advance!

[ebenmichael]

Sorry for the delay here, this should be fixed now!