.png){kind=icon}
.png){kind=icon}
Return
EPL: The event processing language for streaming data
.png){kind=icon}
PRE
AI
DOI：10.1016/j.is.2026.102750.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Stream processing (SP) engines analyse unbounded data with low latency. Among SP languages, the Event Processing Language (EPL) is one of the most mature industrial options, combining SQL-like analytics with Complex Event Processing (CEP). However, EPL lacks a precise formal semantics. This paper provides a denotational semantics for a core EPL fragment covering (i) data model with schema polymorphism and time, (ii) DML including windows, grouping, joins, and (iii) CEP patterns and MATCH_RECOGNIZE. We demonstrate empirical equivalence to Esper (the reference implementation) through a decomposition into atomic queries, which factor non-trivial programs into testable units. The semantics resolves key ambiguities (e.g., scope of NOT, EVERY explosion) and exposes harmful combinations, and we propose a restricted grammar that prevents them. We further formalise window families (time/count, expression-based frames, composite windows) and Allen interval relations for duration events. This work advances the formal foundations of stream languages and aligns EPL with recent formalization efforts in graph and CEP logics.
Keywords:
Event Processing Language
Stream Processing
Denotational Semantics
Complex Event Processing
Window Functions
Journal
I
IF:
3.4
Papers:
109
Citations:
0
View more organizations