Reports: Table CSV Export not formatted properly when using COLLECT(DISTINCT())

[geoffinfosecpersona]

The exported CSV from Neodash differs from the exported CSV from the Neo4j Browser, with the Neo4j Browser export respecting lists as a single column, and the Neodash export treating list entities as individual columns.

In short, it looks like the CSV export process treats commas as delimiters, which causes issues when using a COLLECT(DISTINCT()) on tabular fields, as the lists generated by the COLLECT(DISTINCT()) have the entities in the list separated by commas.

To help us understand your issue, please specify important details, primarily:

• NeoDash version: 2.4.3

• Neo4j Database version: 5.17 Enterprise

• Steps to reproduce

• Create a Tabular Report

• In the report, use a query to return multiple fields that returns multiple properties using COLLECT(DISTINCT)). Sample query using "Movies":

MATCH (n:Movie)
OPTIONAL MATCH (n)-[:ACTED_IN]-(b)
OPTIONAL MATCH (n)-[:DIRECTED]-(c)
OPTIONAL MATCH (n)-[:PRODUCED]-(d)
RETURN 
n.title AS `Movie Name`
, COLLECT(DISTINCT(b.name)) AS `Actors`
, COLLECT(DISTINCT(c.name)) AS `Directors`
, COLLECT(DISTINCT(d.name)) AS `Producers`

• Export the CSV from the tabular report, and the columns won't line up with the exported values (instead of there being 4 columns total, each list entity will create its own column)

• Expected behavior For the above query, there should be 4 columns in the CSV export

• Actual behavior I believe the CSV export should have around 12 columns

[geoffinfosecpersona]

@nielsdejong - Is there a status update for this bug? Thank you!

[nielsdejong]

Hi @geoffinfosecpersona , Thanks for reporting this. To fix this we'll need to make the CSV delimiter configurable as an advanced setting for the table report - you can then choose to go for semicolon or tab, etc.

For now you can probably work around it by using apoc.text.join to turn the list into a ';' seperated string, but it will probably be nicer to customize the CSV delimiter instead.

[geoffinfosecpersona]

Brilliant, @nielsdejong. I'll give that a shot. Thank you!!

Hi @geoffinfosecpersona , Thanks for reporting this. To fix this we'll need to make the CSV delimiter configurable as an advanced setting for the table report - you can then choose to go for semicolon or tab, etc.

For now you can probably work around it by using apoc.text.join to turn the list into a ';' seperated string, but it will probably be nicer to customize the CSV delimiter instead.