Legacy DR10 Unique Source Counts

[John Winterflood]

The Legacy DR10 description page (https://www.legacysurvey.org/dr10/description/) lists the total number for each of the different classificatons of sources (eg PSF, REX, EXP, etc).  However none of these numbers seem to be quite correct.  For instance the total number of all unique sources is given as 2,826,169,461 but a quick SQL query (https://datalab.noirlab.edu/query.php - SELECT COUNT(*) FROM  ls_dr10.tractor_s) indicates a total count of 2,825,807,500 (thus 361,961 sources short).  Is there a known reason for this discrepancy?