Context: I have seen that in version 1.184.0
and later I can add column descriptions within the transform function. [Related question] E.g.:
from transforms.api import transform, Input, Output
@transform(
my_output=Output("/my/output"),
my_input=Input("/my/input"),
)
def my_compute_function(my_input, my_output):
my_output.write_dataframe(
my_input.dataframe(),
column_descriptions={
"col_1": "col 1 description"
}
)
Question: Will this update dataset descriptions only on the branch I run the transformation on?
Yes. You don't need to worry about changing the column descriptions in other branches. The column description only changes the description of the corresponding branch.