Gotchas of Pandas
summary
Shows concrete Pandas pitfalls: NumPy’s lack of NA means ints upcast to float and bools become object when missing values appear, so reindexing often inserts NaNs and changes dtypes; equality checks fail for NaN so use pandas.isnull; reindex is strictly label-based which can produce unexpected NaNs; multithreaded writes (and deletions) can corrupt DataFrames even with locks, especially on large (1–2 GB) sets and when combining with BeautifulSoup/html5 tools; examples target pandas 0.17.
AI-written summary · about
