Sometimes the flow produces a table with totally empty rows (has row ID but no columns). Here we cannot use the Empty Table Switch node unless we use the Remove Empty Rows node first. But this node removes missing value cells (if no other value is available in the row) as well which in some cases is undesired.
How about adding an option to choose between removing and retaining missing cells?
from node description seems to me that Remove Empty Rows node is for removing all rows that contain only missing values so having option to choose between removing and retaining missing cells would be a bit contradictory. Anyways I add @swebb as I got a feeling he knows more about it
Indeed, the node is supposed to remove rows that only have missing values i.e. every column in the row must have a missing value.
Iâm not sure if @armingrudd is saying itâs removing rows when some columns do have values and some donât? If so this is a bug, if not then Iâm not sure I follow.
As a follow up, I just ran a test locally and the node behaves as I expect it to.
This input:
Becomes this output:
The rows that have only missing values have been removed. Rows that have at least 1 value are kept. This is the purpose of the node. Iâm not sure how a dialog option would make sense here.
Could you provide me an expected input and desired output so I can understand a bit better what you want?
Thereâs also the Row Splitter (col+) node which can split a table based on missing values.
This config will put rows with a missing value anywhere in the second output:
I think that @armingrudd means that when he has no columns in the table, he then cannot use the empty row filter on the table to remove all the rows, but even though the table is empty in the sense that it has no cells, it still registers in the Empty Table Switch node as being non-empty because it has rows, which is presumably why he would then like to use the Remove Empty Rows node. My apologies Armin if Iâve got that completely wrong!
The workflow is running automatically. sometimes all columns are filtered. but row IDs remain. So the Empty Table Switch does not work. so I need to remove rows first. But that makes some problems when the table has columns but some rows have missing values which we need to keep for further process. but the Remove Empty Rows node removes them.
I think it means adding an option to choose whether remove missing rows or void rows.
when the table has columns so there would not be any void rows. there the âremove missing rowsâ option would work.
and if there is no columns, there would be no missing values and only void rows. there the âremove void rowsâ would work.
I think we need 2 check box for each. (maybe one needs the void row IDs)
It looks like itâs now available in the nightly/trunk build. If you could let me know if this works as expected Iâll then migrate it over to 3.7, 4.0 and 4.1.
Wouldnât occur on a given input but if you want to remove either instance in a loop a radio button wouldnât let you do that. For example how the node used to work.
But allowing people to select neither may not be appropriate without a warning so maybe that should be added.