If you have large samples, make two separate tables. One for (i) Oxides in %ge and the other (ii) For elements with respect to number of oxygen atoms. If a single sample, you can put it in single Table with separate columns, first column being list of oxides/elements. Anyway, you can always present the actual value 0.00 or 0.02, respectively for MgO and CaO in the Table.
Hien says the data are from EMP analysis, so 3 decimal digits would almost certainly be too many, because the analytical error are typically worse than that (unless you used non-standard analytical conditions). Generally, an EMP output includes a statistical evaluation of the detection limit for each specific oxide for each specific analysis. It is not clear to me if you already have this information, which may or may not have been incorporated in the file you got. In any case, you might still try to go back to the original EMP output and see if this info is given somewhere, if you need it.
Anyway, reporting a value as b.d.l. is always an acceptable and widely used option, provided you know what the detection limit is. This is probably appropriate for your MgO and, possibly CaO. As suggested, reporting 0.00 +/- st.dev is a very good practice, if you analysed several points on the same grain (which would be desirable anyway), because it may give an idea on both the analytical uncertainties and the compositional variability in your grain.
Thank you Paolo Nimis, it is quite difficult to understand, i attached the example, could you explain for me the way i put the data of MgO and CaO into the table of manuscript.
Ok, now I understand that you have individual point analyses showing some 0 or "negative" values (sorry I had missed the minus sign in your post). The negative values are of course meaningless; they simply derive from the correction procedure of the analyses, which is never "perfect". Both negative and 0 values indicate undetectable concentrations. I would simply write 0.00 for all of them.
It is unclear to me which is the unit for the reported detection limits, but I would not bother about this, unless you really need to know if some of the very low reported values (e.g. 0.03 for FeO) are really meaningful. If you just need to show that these values are very low, just report them in the table as they are.
When i wrote 0.00, the supervisor did not agree with me, he ask me to write "< detection limit number". But i do not know how to find it, that is my problem now.
Ok, so you should try to figure out the unit of the detection limits reported in your EMP output. Maybe element ppm? It may well turn out that, e.g., FeO 0.03 should also be reported as "below detection limits".