If I understand correctly, the only thing you are missing to be able to calculate the percentage of events that are special orders, is the total number of events.
I suggest you use appendcols to get that number and then an eval to do the calculation with it afterwards. Like so (where "base search" is your original search from your question):
<base search> | appendcols [search source=order_log | timechart span=1d count as TotalNumOrders] | eval percOfSpecialOrders = (numSpecialOrders/TotalNumOrders) * 100 | table _time, percOfSpecialOrders, percOfTotalPrice
Like stated above. The appendcols will calcualte the daily total number of orders (not just special orders) and append it to each appropriate day from your original search. So your results should look like:
_time,numSpecialOrders,percOfTotalPrice,TotalNumOrders
2014-10-31, 500, 20, 5000
Then you pipe that into the eval to calculate the percentage that are special orders, getting a new column. Looking like this:
_time,numSpecialOrders,percOfTotalPrice,TotalNumOrders,percOfSpecialOrders
2014-10-31, 500, 20, 5000, 10
Finally, since you only care about three fields, you call the table command and list out the fields in the proper order, and voila you have what you wanted.
Hope this helps.
... View more