It would perhaps be an idea to post a sample of your data - obviously dummy values that are confidential but ensure that your sample is an accurate reflection of what you're working with.