Take this dataframe df fragment: (the original is quite larger than this)
df
symbol order_adj symbol_adj 0 NaN NaN NaN 1 USDC