Pandas Interview Questions
Core Concepts
Q: DataFrame operations — select, filter, transform.
python
import pandas as pd
# Create / Load
df = pd.read_csv("data.csv")
df = pd.DataFrame({"name": ["Alice","Bob"], "age": [30, 25]})
# Select
df["name"] # Series
df[["name", "age"]] # DataFrame
df.iloc[0] # Row by integer position
df.loc[df["age"] > 25] # Row by condition
# Filter
df[df["age"] > 25]
df.query("age > 25 and name != 'Bob'")
# Transform
df["age_group"] = df["age"].apply(lambda x: "senior" if x >= 30 else "junior")
df["full_name"] = df["first"] + " " + df["last"]Q: GroupBy and aggregation.
python
# GroupBy: like SQL GROUP BY
df.groupby("department")["salary"].mean()
df.groupby(["dept","level"]).agg({"salary": ["mean","max","count"]})
# Pivot table
pd.pivot_table(df, values="sales", index="region", columns="product", aggfunc="sum")
# Apply custom function
df.groupby("dept").apply(lambda x: x.nlargest(3, "salary")) # Top 3 per deptQ: Merging and joining DataFrames.
python
# Merge (like SQL JOIN)
result = pd.merge(orders, users, left_on="user_id", right_on="id", how="left")
# Concat (stack rows or columns)
combined = pd.concat([df1, df2], ignore_index=True) # Stack rows
wide = pd.concat([df1, df2], axis=1) # Add columnsQ: Performance best practices.
python
# Vectorised operations > apply > loops
# Slow: df.apply(lambda x: x["a"] * x["b"])
# Fast: df["a"] * df["b"] (vectorised)
# Efficient data types
df["category"] = df["category"].astype("category") # String with few unique values
df["int_col"] = df["int_col"].astype("int32") # vs int64
# Chunked reading for large files
for chunk in pd.read_csv("big_file.csv", chunksize=10000):
process(chunk)
# Use numpy where for conditionals
df["flag"] = np.where(df["score"] > 90, "high", "normal")Revision Notes
PANDAS CORE: DataFrame (2D) | Series (1D)
SELECTION: df["col"] | df[["c1","c2"]] | df.loc[condition] | df.iloc[pos]
FILTER: df[df["col"] > value] | df.query("col > value")
GROUPBY: df.groupby("col")["val"].agg({"mean","max","count"})
MERGE: pd.merge(left, right, on="key", how="left/inner/outer")
CONCAT: pd.concat([df1, df2]) = stack rows
PERFORMANCE:
Vectorised ops > apply > loops
astype("category") for low-cardinality strings
Chunked reading for large files
np.where for conditional column creation