File and Object I/O: 10 practice questions

Question 1 of 10

Which pair reads and writes CSV data with pandas?

1. A`read_excel()` and `DataFrame.to_excel()`
2. B`read_parquet()` and `DataFrame.to_parquet()`
3. C`read_json()` and `DataFrame.to_json()`
4. Dpandas `read_csv()` and `DataFrame.to_csv()`

Question 2 of 10

What structure does JSON `orient="records"` produce for a DataFrame?

1. AA nested array containing values without either labeled axis
2. BSeparate arrays for columns, index labels, and data
3. CA list of column-to-value objects without index labels
4. DA mapping from each column to a mapping of index labels

Question 3 of 10

Which worksheet does `read_excel()` read when `sheet_name` is not supplied?

1. AEvery sheet in the workbook
2. BThe first sheet at position zero
3. CNo sheet until a name is supplied
4. DThe last sheet in the workbook

Question 4 of 10

Which `read_csv()` settings directly restrict parsed columns and request their representations?

1. A`index_col` for selection and `skiprows` for requested types
2. B`usecols` for column selection and `dtype` for requested types
3. C`na_values` for column selection and `encoding` for requested types
4. D`header` for selection and `quotechar` for requested types

Question 5 of 10

Which JSON orientation keeps `columns`, `index`, and `data` in separate entries?

1. AThe `split` orientation
2. BThe `values` orientation
3. CThe `columns` orientation
4. DThe `records` orientation

Question 6 of 10

What does `read_excel(..., sheet_name=["North", 3])` return?

1. AA dictionary containing the named sheet and sheet at position three
2. BOnly the first sheet because lists are ignored by `sheet_name`
3. COne DataFrame formed by vertically appending every workbook sheet
4. DA list of cell values with all sheet labels removed

Question 7 of 10

How can a Parquet reader avoid materializing unneeded columns?

1. APass the required names through `columns`
2. BPass the required names through `partition_cols`
3. CPass the required names through `compression`
4. DPass the required names through `index`

Question 8 of 10

A pipe-delimited file has no header, `NA?` is its sole missing marker, and `account_id` must remain text. Only three named columns are needed. Which ingestion design satisfies all constraints?

1. ASet `sep="|"`, explicit `names`, `header=None`, `usecols`, text dtype for `account_id`, `na_values=["NA?"]`, and `keep_default_na=False`
2. BUse `read_json(orient="records")` and treat each pipe-delimited line as a JSON object
3. CSet `header=0`, keep every column, and rely on the default missing-marker set
4. DUse default `read_csv()` inference, then rename columns and convert all values to floats

Question 9 of 10

A DataFrame must round-trip through JSON while preserving dtypes and the index name, and the reader must reconstruct the same orientation. Which approach is supported?

1. AWrite `values` and read with the default orientation
2. BWrite with `orient="columns"` and read with `orient="split"`
3. CUse `orient="table"` for both the write and the read
4. DWrite with `orient="records"` and infer the omitted index name later

Question 10 of 10

A DataFrame with a custom index will be shared with a non-pandas consumer that expects exactly the named data columns. The file must use Parquet and the consumer rejects unexpected index fields. What should the writer do?

1. ACall `to_parquet(..., index=False)` and accept that the custom index will not round-trip
2. BCall `to_parquet(..., index=True)` so the consumer receives no index column
3. CRely on each engine's default index behavior and promise an identical schema
4. DRename the index to a non-string label because Parquet requires such names