if the dataset fits into memory, the easiest is to use python, for something like
df = dataiku.Dataset("the_dataset_name").get_dataframe()
m = df["the_column_name"].median()
For larger datasets, you can make a Window recipe, set the window to be ordered on the column, activate the window frame but leave limit preceding and limit following unchecked, tick the "cumulative distreibution" aggregate, and filter + use a TopN recipe to get the first row with a cumedist value above 0.5
If the question is about enriching each row with the median of the column, then you'll have to sync the data to a SQL database and use SQL to perform the operation.