md` ### Pattern String Syntax Pattern strings provide a simple human-readable way to encode sequence patterns. Their syntax is inspired by [pattern encoding in the graph query language Cypher](https://neo4j.com/docs/cypher-manual/current/syntax/patterns/) and [regular expressions](https://en.wikipedia.org/wiki/Regular_expression) but is much simpler to learn and use. Here is the syntax: **"()"** - matches any one event **"(:name)"** - matches one event called "name" ("_eventName" property) **"(:name1|name2)"** - matches one event called "name1" OR "name2" **"(!:name)"** - matches one event NOT called "name" **"({prop1=propName, prop2>=3, prop3!=1|2})"** - matches one event with properties: "prop1" equal to "propName", "prop2" greater or equal 3 and "prop3" not equal 1 or 2. Supported operators include: "==", "=" and ":" (all represent "equal"), "!=", ">", "<", ">=" and "<=". **"[3]\\()"** - matches any event subsequence of length 3 **"[0..5]\\()"** - matches any event subsequence of length 0 to 5 events **"[2..]\\()"** - matches any event subsequence of length 2 or more events **"[..4]\\()"** - matches any event subsequence of length 0 to 4 events **"[..]\\()"** - matches any event subsequence of any length **"[2..3]"** - matches any event subsequences of length 2 to 3 events Basically, an event pattern consists of "[...]" part encoding the possbile count of events followed by "(...)" part encoding the constraints of each event. If "[...]" part is missing, it defaults to one event ("[1]"). If "(...)" part is missing, it defaults to any event ("()"). All the above rules can be combined IN THE ORDER INTRODUCED ABOVE, for example: "[..5]\\(!:name1|name2 {prop1=5|6, prop2!=propName1|propName2})". To match a sequence of different events, separate them with one or more "-" symbols: **"(:name1)-(:name2)"** - matches one event called "name1" immediately followed by one event called "name2" **"[1..2]\\(:name1|name2)--[2]\\(:name3)"** - matches 1 or 2 events called "name1" or "name2" immediately followed by 2 events called "name3" **"(:name1)-[..3]-(:name2)"** - matches one event called "name1" later followed by one event called "name2" with any events in between but no more than 3 of them To match the start or end of a sequence, use "|" symbol: **"|-(:name)-..."** - matches one event called "name" at the start of a sequence **"...-(:name)-|"** - matches one event called "name" at the end of a sequence Omitting "|" at the start or end of a pattern string defaults to matching any number of any events in a sequence before or after, respectively, the explicitly matched sub-sequence. So "-(:name1)-(:name2)" pattern string matches any sequences, which have one event called "name1" immediately followed by one event called "name2" ANYWHERE inside of it. There are a few rules for specifying whether a match should represent a whole sequence or only some part of it. They are handy for extracting and modifying sub-sequences via extractPattern() and modifyPattern() functions, respectively. Here is the syntax: **"-(:name)-..."** - starts a match from the start of a sequence (equivalent to "|-[..]-(:name)-...") **"(:name)-..."** - starts a match from event called "name" but a matching sequence can have other events before it **"-(:name1)-(^:name2)-..."** - starts a match from event called "name2" **"^-(:name1)-..."** - starts a match from the start of a sequence **"...-($:name1)-(:name2)-..."** - ends a match at event called "name1" **"...-(:name1)-$"** - ends a match at the end of a sequence **"-(:name1)-[1..3]\\(^$:name2)-(:name3)-..."** - collapses a match to only a 1-3 event sub-sequence of events called "name2" Finally, it is possible to specify meta-parameters and sequence-level properties with a "<...>" part, which needs to be placed at the end of a pattern string after an event pattern part: **"...<_whereToMatch=_ALL>"** - extracts ALL matches in a sequence, so it can split a sequence into several sub-sequences. Default value is "_FIRST", which extracts only the first match for each sequence. **"...<_allowOverlaps=true>"** - extracts all overlapping matches in a sequence (if "_whereToMatch" equals "_ALL"). Default value is "false", which starts the next match within a sequence after the end of a previous match. **"...< seqProp1=name1|name2, seqProp2!=3 >"** - matches sequences with sequence-level properties: "seqProp1" equal to "name1" or "name2" and "seqProp2" not equal to 3 All SPACES in a pattern strings are ignored, so you can add them as you please for better readability. Pattern string syntax is easy to master, apply and read; nevertheless, it provides a powerful interface to express almost any desired sequence question. `