(This answer is based on experience with the business side of sales forecasting, more so than on rigorous statistical/mathematical knowledge)
Looking at your data, it makes more sense to forecast it at a weekly level than at a daily level. At at daily level it is too sparse, but at a weekly level you would have a more meaningful times series.
week 1: 0,0,1,0,0,0,0
week 2: 2,0,0,0,0,0,0
week 3: 0,0,0,1,0,0,0
week 4: 1,0,1,0,0,0,0
week 5: 0,0,0,0,0,0,0
week 6: 1,0,0,2,0,0,0
Any forecasting method you would use at a daily level, would give a fractional value per day. This doesn't really help, since these are sales units, so a forecast value of ~ 0.14 doesn't mean much, unless you interpret it as a probability (and I don't know enough math to help in that case, but others might know better how to treat that).
If you aggregate the data by week, you get:
week 1: 1
week 2: 2
week 3: 1
week 4: 2
week 5: 0
week 6: 3
You can then simply average that value over all the weeks you have, or maybe use a moving average. You would then get an average of 3 units sold per two weeks.
Keep in mind that this is a sales forecast: What is the purpose of a sales forecast? To make sure that you have enough inventory to satisfy customers' demand. Based on the method I described above, you would know that you need to ship/order 3 units of inventory every 2 weeks to satisfy the demand for that product - without going into ARIMA or Exponential smoothing or some other more involved time series analysis.