I see the change makes sense On Wed, Jan 27, 2016 at 10:13 PM, Priyanka Gugale <[email protected]> wrote:
> The startRow purpose is as you explained, but the value assigned to it is > not serving the right purpose. As per my understanding the value should be > value of primary key of start row. I had to read through code to understand > that. A meaningful name and right javadoc will make it easy to understand. > > -Priyanka > > On Wed, Jan 27, 2016 at 9:40 PM, Pramod Immaneni <[email protected]> > wrote: > > > Priyanka the start in the query should act like a cursor and keep > > getting updated when new data is pulled in and that's possible the > > reason for that name. Is that not happening? > > > > > On Jan 27, 2016, at 3:53 AM, Priyanka Gugale <[email protected]> > wrote: > > > > > > Hi, > > > > > > CassandraPOJOInputOpeartor in Malhar accepts query in the form: > > > "select * from %t where token(%p) > %s limit %l" > > > Where %s is start value. > > > > > > *What is "start value" here?* > > > From javadoc in code it looks like %s is the row number in table from > > where > > > you want to start scan. But looking at code it should be > > "primaryKeyValue" > > > of row from where you want to start scan. (I didn't find any way in cql > > to > > > scan starting from some row number, unless you have rowNumber column in > > > your table.) > > > > > > If this understanding is correct we should rename this to something > like > > > startRowPrimaryKey and update javadoc. > > > > > > -Priyanka > > >
