changed the processing order - now: spel, regex, filters

core/src/main/java/com/redislabs/riot/KeyValueProcessingOptions.java

@@ -18,20 +18,17 @@ public class KeyValueProcessingOptions {
 
     @Option(arity = "1..*", names = "--spel", description = "SpEL expression to produce a field", paramLabel = "<f=exp>")
     private Map<String, String> spelFields = new HashMap<>();
-    @Option(arity = "1..*", names = "--filter", description = "SpEL expression to filter records", paramLabel = "<exp>")
-    private List<String> filters = new ArrayList<>();
     @Option(arity = "1..*", names = "--var", description = "Register a variable in the SpEL processor context", paramLabel = "<v=exp>")
     private Map<String, String> variables = new HashMap<>();
     @Option(names = "--date", description = "Processor date format (default: ${DEFAULT-VALUE})", paramLabel = "<string>")
     private String dateFormat = new SimpleDateFormat().toPattern();
     @Option(arity = "1..*", names = "--regex", description = "Extract named values from source field using regex", paramLabel = "<f=exp>")
     private Map<String, String> regexes = new HashMap<>();
+    @Option(arity = "1..*", names = "--filter", description = "SpEL expression to filter records", paramLabel = "<exp>")
+    private List<String> filters = new ArrayList<>();
 
     public ItemProcessor<Map<String, Object>, Map<String, Object>> processor(StatefulConnection<String, String> connection) {
         List<ItemProcessor<Map<String, Object>, Map<String, Object>>> processors = new ArrayList<>();
-        if (!filters.isEmpty()) {
-            processors.add(new FilteringProcessor(filters));
-        }
         if (!spelFields.isEmpty()) {
             processors.add(new SpelProcessor(connection, new SimpleDateFormat(dateFormat), variables, spelFields));
         }
@@ -40,6 +37,9 @@ public ItemProcessor<Map<String, Object>, Map<String, Object>> processor(Statefu
             regexes.forEach((f, r) -> fields.put(f, RegexNamedGroupsExtractor.builder().regex(r).build()));
             processors.add(new MapProcessor(fields));
         }
+        if (!filters.isEmpty()) {
+            processors.add(new FilteringProcessor(filters));
+        }
         if (processors.isEmpty()) {
             return null;
         }

docs/processing.adoc

@@ -1,31 +1,44 @@
 The following processors can be applied to records in that order:
 
-* Filters
 * Transforms
 * Regular expressions
-
-==== Filter
-
-Keep records that match a https://docs.spring.io/spring/docs/current/spring-framework-reference/core.html#expressions[Spring Expression Language] (SpEL) boolean expression
-
-`--filter "name matches '[a-zA-Z\\s]+'"`
+* Filters
 
 ==== Transform
 
-Produce key/value pairs using SpEL: `field1=<exp>`, `field2=<exp>`, ...
+Produce key/value pairs using https://docs.spring.io/spring/docs/current/spring-framework-reference/core.html#expressions[Spring Expression Language] (SpEL): `field1=<exp>`, `field2=<exp>`, ...
 
-For example `--spel field1=' generates a field named `field1` with always the same value `foo`.
+For example `--spel "field1='foo'"` generates a field named `field1` with always the same value `foo`.
 
 Input fields are accessed by name (e.g. `field3=field1+field2`).
 
 The processor also exposes the following variables that can be called with the `#` prefix:
 
-* `redis` : Redis connection to issue any command, e.g. `name=#redis.hgetall('person1').lastName`
-* `date` : date parser/formatter, e.g. `epoch=#date.parse(mydate).getTime()`
-* `index` : sequence number of the item being generated, e.g. `id=#index`
+===== redis
+Redis connection to issue any command:
+
+`--spel "name=#redis.hgetall('person1').lastName"`
+
+===== date
+Date parser/formatter:
+
+`--spel "epoch=#date.parse(mydate).getTime()"`
+
+===== index
+Sequence number of the item being generated:
+
+`--spel "id=#index"`
 
 ==== Regex
 
 Extract patterns from source fields using regular expressions:
 
-`--regex "name=(?<first>\w+)\/(?<last>\w+)"`
+`--regex "name=(?<first>\w+)\/(?<last>\w+)"`
+
+==== Filter
+
+Keep records that match a SpEL boolean expression.
+
+`--filter "value matches '\\d+'"` will only keep records where the `value` field is a series of digits.
+
+