Constructing a unique identifier for each row in lsst_sim

Question

Hi datalab, I'd like to make a unique ID per object in the lsst_sim.simdr2 catalog, for example by combining values of existing columns into a string. The first idea was to combine ring256 and random_id, but that's not always unique in high-density areas. Is there some set of columns that could be combined to provide a unique string, or a better way of getting unique identifiers? Thanks!

Answer 1

Hi Claire, thanks for reaching out. Maybe you could try encoding RA and Dec (possibly in conjunction with other fields) into a unique ID? Both RA and Dec are DOUBLE PRECISION, so hopefully have enough significant digits to be unique even in dense regions.

Another option would be to simply enumerate all objects (with a BIGINT). Not sure how expensive such an operation would be though...

Robert

Comments

Hi Robert, thanks for the suggestion.
Would you consider adding a BIGINT column to the catalog in the datalab? It seems like it might be useful to others as well..
Claire