I just tried the following with the neo4j web interface:
I wrote a string of 26 characters and duplicated it through 32,000 lines, which is a total of 832,000 characters.
I created a node with the text property and copied my text in it, and it worked perfectly.
I tried again with 64,000 lines with spaces at the end of lines, for a total of 1,728,000 characters. Created a new node, then requested node and copied the result back to the file to check the size (you never know), and wc gave me 1728001 (the error should be in the copy / paste process, I suppose).
He did not seem to complain.
FYI is equivalent to a text with 345600 words of average size 4 and a space (5 characters) and a book of 1000 pages with 300 words per page.
I do not know how this can affect performance if there are too many nodes. If this doesn’t work because of this, you can always think that neo4j stores relationship information, with a property identifier as an identifier for another document-oriented database for text extraction (or just a file path as a property path) .
source share