Too many levels within factor for Anova?

I'm having trouble carrying out analysis of my data. I'm trying to test whether the 'median' value of pollution has an effect on the 'yield' of a crop in relation to the 'state' and 'year'. There are 8 states and data for years 2016-2018 with some states having data for 2015 resulting in 27 levels for a factor. Is a 2-way Anova the right method or are there too many levels within the factors?

Data table summary of data

question from:https://stackoverflow.com/questions/65830105/too-many-levels-within-factor-for-anova

与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…