Well, unfortunately my crystal ball is broken so I can only speculate unless you can tell us that that 9.0 reading was.
If that was a regular blood test, which you can, in principle do without fasting, then 9.0 mmol/l is a high but not high enough to diagnose diabetes - diagnostic cut-off for random BG test is >11.1 mmol/l.
A random blood test can be used to diagnose diabetes, but not exclude it, so it makes some sense to send you for a OGTT.
A negative OGTT means you're probably not diabetic (but might have impaired fasting glucose), which would explain what the 2nd doctor said [for example, you having a bottle of Lucozade before the first test might have caused this but since I don't know any details I can only speculate]
Another possibility is that the first test was an HbA1c, in which case 9.0 is probably a rather high 9.0% (DCCT, equivalent to 75 mmol/l in new IFCC units) - but then again, they'd have told you that the HbA1c is 75 because they are supposed to use the new IFCC units now. An HbA1c of 9.0% is probably high enough to diagnose diabetes (WHO recommends >6.5% for diagnosing diabetes, not sure what the NICE guidelines say on this). In this case, it seems very unlikely that the OGTT would be "normal" (that HbA1c corresponds to an average blood glucose of 13 mmol/l) - in that case, I'd ask for the detailed test results, and would ask for a fasting blood test (pharmacies might do diabetes screening - you could try skipping breakfast and going for test 1st thing in the morning if you GP doesn't want to do another test)
Again, the above is pure speculation. I'd recommend that you find out what exactly that "9.0" test was, because then we might be able to give you more specific advice.