cancel
Showing results for 
Search instead for 
Did you mean: 

Unexpected list index behavior

tms
Graph Buddy

I am seeing unexpected list index behavior while using negative indices.

Here is a trivial list:

 

WITH
    RANGE(1,10) AS indices
RETURN
    indices

 

I would like to answer the last item of the list as well as the last three items.

I therefore use the following cypher:

 

WITH
    RANGE(1,10) AS indices
WITH
    indices[-1] AS lastIndex,
    indices[-4..-1] AS lastThreeIndices
RETURN
    lastIndex,
    lastThreeIndices

 

When I run this in the Neo4J Browser, I see an answer of  10, as expected, for lastIndex.

I get the following as the value of lastThreeIndices:

 

[7, 8, 9]

 

I expect to see

 

[8, 9, 10]

 

If indices[-1] gives the correct answer (the last element of the list), then how can it be possible for indices[-4..-1] to NOT answer the last three elements of the list?

Am I having a brain-fart? What am I missing?

This is at least unexpected if not simply a bug.

1 ACCEPTED SOLUTION

bennu_neo
Neo4j
Neo4j

Hi @tms ,

You are just missing the fact that indexing and slicing are two different things on list comprehensions.

Take a look on https://towardsdatascience.com/mastering-indexing-and-slicing-in-python-443e23457125

You want something like:

WITH
    RANGE(1,10) AS indices
WITH
    indices[-1] AS lastIndex,
    indices[-3..] AS lastThreeIndices
RETURN
    lastIndex,
    lastThreeIndices
Oh, y’all wanted a twist, ey?

View solution in original post

2 REPLIES 2

bennu_neo
Neo4j
Neo4j

Hi @tms ,

You are just missing the fact that indexing and slicing are two different things on list comprehensions.

Take a look on https://towardsdatascience.com/mastering-indexing-and-slicing-in-python-443e23457125

You want something like:

WITH
    RANGE(1,10) AS indices
WITH
    indices[-1] AS lastIndex,
    indices[-3..] AS lastThreeIndices
RETURN
    lastIndex,
    lastThreeIndices
Oh, y’all wanted a twist, ey?

I missed that '..' is Cypher's way of spelling the Python ':' (slice) operator. 🙂

That works, I appreciate the quick response.