这个限制不仅仅在janino这边，我们这边已经尝试去除了janino这边的64kb的限制，但是逻辑运行到jdk这边也出现了同样的限制。目前spark是通过spark.conf.set("spark.sql.codegen.wholeStage",false)来避免类似的问题，所以我觉得最好的方法是可以通过配置来限制一些SQL逻辑计划的优化，避免不同的view的逻辑合并到一起，这样就能通过将拆分sql来将逻辑拆分到不同的function中，从而避免一个类函数太大。Caused by: java.lang.ClassFormatError: Invalid method Code length 123732 in class file ExpressionReducer$3674 at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:763) at org.codehaus.janino.ByteArrayClassLoader.findClass(ByteArrayClassLoader.java:77) at java.lang.ClassLoader.loadClass(ClassLoader.java:424) at java.lang.ClassLoader.loadClass(ClassLoader.java:357) at org.apache.flink.table.runtime.generated.CompileUtils.doCompile(CompileUtils.java:86) at org.apache.flink.table.runtime.generated.CompileUtils.lambda$compile$1(CompileUtils.java:66) at org.apache.flink.shaded.guava18.com.google.common.cache.LocalCache$LocalManualCache$1.load(LocalCache.java:4742) at org.apache.flink.shaded.guava18.com.google.common.cache.LocalCache$LoadingValueReference.loadFuture(LocalCache.java:3527) at org.apache.flink.shaded.guava18.com.google.common.cache.LocalCache$Segment.loadSync(LocalCache.java:2319) at org.apache.flink.shaded.guava18.com.google.common.cache.LocalCache$Segment.lockedGetOrLoad(LocalCache.java:2282) at org.apache.flink.shaded.guava18.com.google.common.cache.LocalCache$Segment.get(LocalCache.java:2197) ... 38 more -- Sent from: http://apache-flink.147419.n8.nabble.com/
