The　social　networking,　network　monitoring　and　financial　applications　have　a　need　to　query　high　rate　streaming　of　XML　data,　but　previous　methods　for　executing　XPath　queries　on　streaming　XML　data　have　not　kept　pace　with　multicore　CPUs.　Data　parallel　query　methods　can　improve　processing　efficiency　by　using　multi-core　CPU.　Push　down　transducer　as　a　special　type　of　automaton,　it　not　only　can　be　used　to　handle　XML　data,　and　its　internal　stack　structure　can　preserve　the　important　information　in　the　processing,　so　it　can　combine　with　the　data　parallel　querying　to　improve　the　processing　efficiency　further.　In　order　to　improve　the　efficiency,　transfers　XPath　query　pattern　into　a　series　of　ordered　subquery　which　its　result　are　valid　and　has　easier　logic.