Andries, I ended up with this. I just don't know why some things don't exist. Some, I think are new but others such as being able to read strings as json from a database would be really helpful. Being able to read in Json, Hstore, and XML from PostgreSQL like it were on the file system would be great. Thanks for the help. The tool is really beneficial in lowering maintenance time costs for ETL.
-Andrew Evans ________________________________________ From: Andries Engelbrecht [[email protected]] Sent: Wednesday, May 25, 2016 3:47 PM To: [email protected] Subject: Re: One more with Excepting Columns by name Perhaps look into using a predicate with typeof. --Andries > On May 25, 2016, at 11:15 AM, Andrew Evans <[email protected]> > wrote: > > Drill Community, > > Thanks for the quick response to my last question. I am exploring the tool > and will definitely use it, especially the json capabilities with the flatten > function. I noticed that Flatten will error on nested maps. I would like to > exclude these maps to build tables from them later by excluding them since I > do not know every key in the Json mapping. Is there a way to select * keys > from a json object except the keys that are maps? > > The problem is that I would need to read through millions of records tens to > hundreds of times per day to achieve productivity with this method. I really > want to avoid anything manual. > > I am looking for something like: > > SELECT flatten(filter_not(record[column],array['col1','col2'])) FROM table > > > All help is greatly appreciated. > > Thanks, > > Andrew Evans
