A café logs every milky drink it sells: the cup size and the milk. Here are Saturday morning's 20 orders, stored in a DataFrame called orders (read down the left half, then down the right):
| order_id | size | milk | order_id | size | milk | |
|---|---|---|---|---|---|---|
| 101 | large | oat | 111 | large | oat | |
| 102 | small | dairy | 112 | medium | dairy | |
| 103 | medium | oat | 113 | small | soy | |
| 104 | large | dairy | 114 | large | dairy | |
| 105 | large | oat | 115 | medium | oat | |
| 106 | small | oat | 116 | small | dairy | |
| 107 | medium | soy | 117 | large | oat | |
| 108 | large | soy | 118 | medium | dairy | |
| 109 | small | dairy | 119 | large | soy | |
| 110 | medium | oat | 120 | small | oat |
The owner suspects that people who order oat milk go large more often than people who order dairy. To check, she counts the orders in a grid, one row per size and one column per milk, and compares two shares:
What does the last line print? The code keeps full precision and rounds to 3 decimal places only at the very end, so do the same.