Why does a load job load table headers as nodes into the database

[kwonder]

This is part of my schema definition：

My load job is defined as follows:

My GenFile looks like this：

However, when I run the load, the edge HAS_GENERATOR will use the header as data，So we have these two extra nodes.

This problem has been bothering me for two days and I haven’t been able to figure out why this is happening because I already used USING header="true". And that doesn’t happen with any other nodes or edges.

I noticed the following strange phenomenon in the load log, which did not treat the table header as an invalid row:

I tried using VERTEX_MUST_EXIST="true", but unfortunately this statement didn’t work

I would appreciate it if anyone could tell me what to do

[Jon_Herke]

@kwonder that does seem like a bug (doesn’t work as expected). Our team will look into replicating the issue. If it is indeed a pyTigerGraph issue we will issue a fix and release a new version of pyTigerGraph. Thank you for the very detailed step by step walkthrough!