Amazon-S3 translator does not support glob based pattern like *.json to select the all files in a Bucket. However, this translator gives you a procedure called "list()" and table based on this procedure is called "Bucket". So in effect, you can get the list of files using "select * from Bucket", then using that you can read each file (or just updated files) in virtual procedure perhaps.

Ramesh..

If you got the above example to work, then it is the process of making it work for many files. For that I am saying to follow the something like below pseudo code

create virtual procedure myrecords () returns table (x, y, z)
begin
  loop on (select key from bucket) as files
  begin
      insert into #records FROM SELECT SP.x as x, SP.y as y, SP.z as z  FROM (EXEC s3.getTextFile(name=>files.key)) AS f,  TEXTTABLE(f.file COLUMNS e1 integer, e2 string, e3 double HEADER) AS SP;
  end
  select * from #records
end

Note the above is written very loosely, you need to correct it for reading JSON files, return format etc. Above gives an idea what I was thinking about, do not expect to work as is.

Ramesh..