Open GoogleCodeExporter opened 8 years ago
Original comment by sjaco...@ucr.edu
on 19 May 2015 at 10:47
Just discovered that this failure is sporadic, and that it actually produces a
different optimized Algebricks plan in the times that it fails. Is it supposed
to be possible for the optimized plan to be nondeterministic?
The difference in the plan is here:
When it works we have
project ([$$63])
-- STREAM_PROJECT |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
unnest-map [$$62, $$63] <- function-call: asterix:index-search, Args:[AString: {times}, AInt32: {0}, AString: {emergencyTest}, AString: {CHPReports}, ABoolean: {false}, ABoolean: {false}, ABoolean: {false}, AInt32: {1}, %0->$$61, AInt32: {0}, TRUE, TRUE, FALSE]
-- BTREE_SEARCH |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
assign [$$61] <- [ADateTime: { 2015-05-20T21:42:19.365Z }]
-- ASSIGN |PARTITIONED|
empty-tuple-source
-- EMPTY_TUPLE_SOURCE |PARTITIONED|
When it fails we have:
project ([$$63])
-- STREAM_PROJECT |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
unnest-map [$$62, $$63] <- function-call: asterix:index-search, Args:[AString: {times}, AInt32: {0}, AString: {emergencyTest}, AString: {CHPReports}, ABoolean: {false}, ABoolean: {false}, ABoolean: {false}, AInt32: {1}, %0->$$61, AInt32: {0}, TRUE, TRUE, FALSE]
-- BTREE_SEARCH |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
replicate
-- SPLIT |PARTITIONED|
exchange
-- ONE_TO_ONE_EXCHANGE |PARTITIONED|
assign [$$61] <- [ADateTime: { 2015-05-20T21:44:31.633Z }]
-- ASSIGN |PARTITIONED|
empty-tuple-source
-- EMPTY_TUPLE_SOURCE |PARTITIONED|
So sometimes a replicate exchange happens and sometimes not...
Original comment by sjaco...@ucr.edu
on 20 May 2015 at 9:59
Original issue reported on code.google.com by
sjaco...@ucr.edu
on 19 May 2015 at 10:46