inconsistent output with large? datasets!!! please see this

[dxl3742]

Hi, I have a dataset with 344393 clonotypes, when I upload this into VDJdb it gives me around 500 matches, but when I split the same dataset into 3 parts (~10000 each) each individual gives me around ~10000 matches under the same search conditions. I have also used the java VDJmatch executable with the same conditions and still encountering this problem.