Neo4j: Split string and get position

问题:

I need to split a field in different values and store each value in a different node. For each created node I want to store the position. Example:

Sentence             Words
My car is red        My;car;is;red

Using:

FOREACH (w IN SPLIT(line.TWords, ";") |
 MERGE (wd:Word {word: w})

I can split the field and store the different words, but I'd like to store the position on the relationship.

My car is red  -[HAS_WORD {position:1}]-> My
My car is red  -[HAS_WORD {position:2}]-> car
My car is red  -[HAS_WORD {position:3}]-> is
My car is red  -[HAS_WORD {position:4}]-> red

How can I get this?

SOLUTION

USING PERIODIC COMMIT LOAD CSV WITH HEADERS  FROM 'file:///output_2016-05-06_0203_Neo4jImport.csv' AS line FIELDTERMINATOR "\t"

MERGE (s:Segment{text: line.Source})
MERGE (ta:Segment{text: line.Target})

WITH SPLIT(line.SWords, ";") AS SWords, line, s, ta
UNWIND RANGE(0, SIZE(SWords)-1) as i
    MERGE (s)-[r:HAS_WORD {position:i+1}]->(w:Word {word: SWords[i]})

WITH SPLIT(line.TWords, ";") AS TWords, line, ta
UNWIND RANGE(0, SIZE(TWords)-1) as i
    MERGE (ta)-[r:HAS_WORD {position:i+1}]->(w:Word {word: TWords[i]})

Be sure that the fist WITH has the variable references necessary in second WITH: WITH SPLIT(line.SWords, ";") AS SWords, line, s, ta

回答1:

You can use a range based on the size of the split, assuming the node containing the sentence is identified with sentence :

WITH split(line.TWords, ';') as splitted
UNWIND range(0, size(splitted) -1) as i
MERGE (w:Word {word: splitted[i]})
MERGE (sentence)-[:HAS_WORD {position: i}]->(w)

Update

USING PERIODIC COMMIT LOAD CSV WITH HEADERS 
FROM 'file:///output_2016-05-06_0203_Neo4jImport.csv' 
AS line FIELDTERMINATOR "\t" 
MERGE (s:Segment{text: line.Source}) 
WITH SPLIT(line.SWords, ";") AS SWords, line 
UNWIND RANGE(0, SIZE(SWords)-1) as i 
MERGE (s)-[r:HAS_WORD {position:i+1}]->(w:Word {word: SWords[i]}) 

回答2:

Use range:

MERGE (S:Sentence {text:"My car is red"})
  WITH S, SPLIT(S.text, " ") as words
  UNWIND RANGE(0,SIZE(words)-1) as i
    MERGE (S)-[r:HAS_WORD {position:i+1}]->(w:Word {word: words[i]})
RETURN S, r, w