Possible with repeated plate-M# at different dates

[murraycutforth]

I wanted to check the number of rows in the database corresponding to each plate - measurement number, and this is what I found:

code: print(df.groupby(["plate", "measurement"]).size())

plate  measurement
2      M1             383
       M3             383
       M4             383
       M6             383
3      M1             383
       M2             383
       M3             383
       M4             383
       M6             383
4      M2             383
       M3             383
       M4             383
       M5             383
       M6             383
5      M1             383
       M2             383
       M3             383
       M4             383
       M5             383
       M6             383
6      M2             383
       M3             383
       M4             383
       M5             383
       M6             383
7      M1             383
       M2             383
       M3             383
       M4             383
       M5             383
       M6             383
9      M1             383
       M2             383
       M3             383
       M5             383
       M6             383
10     M1             383
       M2             383
       M3             383
       M4             383
       M5             383
99     M1             768
       M2             384
       M3             384
       M4             384
       M5             768
       M6             384

A few points stand out. I think it would be simpler to treat A1 the same as other wells, and insert a row with all "NaN"s so all the counts end up at 384, since in plate 99 A1 is not blank. Secondly, 99-M1 and 99-M5 have been repeated at different dates, leading to this surprising number. We should delete the older version if that data is being thrown away, but I'm not sure if the old data is still meaningful, just wanted to draw attention.