Nested data structures — list of dicts, dict of lists: Visual walkthrough

1.2.26 · D2 · Coding › Introduction to Programming (Python) › Nested data structures — list of dicts, dict of lists

Shuru karne se pehle, vocabulary ke baare mein ek promise. Hum exactly yeh plain-English words use karenge aur har ek ko picture par earn karenge:

ek cell = ek single fact (Asha ki marks: 88),
ek row = ek cheez ke baare mein saare facts (Asha ke baare mein sab kuch),
ek column = saari cheezon mein ek tarah ka fact (sabki marks).

Jinpe hum lean karte hain wo prerequisites hain: Lists (ek ordered shelf jise tum position 0,1,2,… se index karte ho), Dictionaries (ek labelled drawer jise tum name/key se kholte ho), aur List Comprehensions (ek compact sentence jo ek list build karti hai). Agar inme se koi shaky lagta ho, pehle unhe open karo.

Step 1 — Ek bare grid se shuru karo (abhi tak koi Python nahi)

KYA. Code bhool jao. Data ko ek plain table ki tarah draw karo: names ek side mein, facts cells mein.

KYUN. Jo do structures hum banane wale hain dono isi exact grid ko store karte hain. Agar hum pehle grid ko anchor kar lein, toh baad ki har shape bas "main is grid pe kaise chalu?" hai — yaad karne ke liye kuch naya nahi.

PICTURE. Teen cheezein (Asha, Ravi, Meera), har ek ke paas teen facts hain (name, age, marks). Red cell wo ek fact hai jise hum har diagram mein chase karenge: Ravi ki marks = 73.

Figure — Nested data structures — list of dicts, dict of lists

Step 2 — Grid ko ROWS mein slice karo → List of Dicts

KYA. Grid ko horizontal cuts se kaato. Har horizontal strip ek poori cheez hai. Har strip ko ek labelled drawer mein pack karo — ek dict {column: value} — aur drawers ko ek shelf par line up karo — ek list.

KYUN horizontal? Kyunki sabse common real sawaal yeh hota hai "mujhe ek record ke baare mein sab kuch do" (ek API result, ek form submission). Horizontal strips ise ek single grab bana deti hain.

PICTURE. Red cell 73 ko dekho: yeh second drawer (Ravi ke) ke andar land hota hai, label marks ke neeche.

Step 3 — SAME grid ko COLUMNS mein slice karo → Dict of Lists

KYA. Ab usi grid ko vertical cuts se kaato. Har vertical strip sabke liye ek tarah ka fact hai. Har strip ko ek naam do (marks) aur strip ko ek list ke roop mein store karo, phir un named lists ko ek dict mein rakho.

KYUN vertical? Kyunki doosra common sawaal yeh hota hai "ek poore column par maths karo" (sabki marks ka average). Ek vertical strip pehle se hi numbers ki flat list hai — sum ke liye ready.

PICTURE. Red 73 ab marks list mein rehta hai, position 1 par. Data ke baare mein kuch nahi badla — sirf cuts ki direction badli.

Step 4 — Dono readings ko side by side rakho: index crosses over karta hai

KYA. Bilkul same red cell ke liye dono formulas ko line up karo aur ghoor ke dekho.

KYUN. Yahi parent note ka poora point hai. Symbols ko positions swap karte dekhna yahi "transpose" ka matlab hai — ek vague analogy nahi, same do coordinates ka ek literal reordering.

PICTURE. Red arrow dikhata hai i aur c LoD se DoL jaate waqt seats trade karte hain.

Step 5 — Column maths: KYUN DoL jeet jaata hai (average derive karo)

KYA. Dono shapes mein average marks compute karo aur kaam gino.

KYUN. Parent claim karta hai ki DoL column maths ke liye better hai. Aao dekhte hain ki LoD kitna extra kaam force karta hai.

PICTURE. Left par (LoD) marks teen drawers mein bikre hain — red loop ko har drawer visit karna hoga aur ek fact bahar nikalna hoga (ek gather). Right par (DoL) marks pehle se hi ek single red strip hain — sum directly chalti hai, bina gathering ke.

Worked example Dono shapes mein average marks

# LoD — must GATHER first
total = sum(s["marks"] for s in students)   # visit every drawer, pull "marks"
avg   = total / len(students)               # len = number of rows
 
# DoL — column is already flat
avg   = sum(students["marks"]) / len(students["marks"])

Numbers 88, 73, 95 ke saath: $\frac{88 + 73 + 95}{3} = \frac{256}{3} = 85.33 \overline{3}$ . LoD ko column collect karne ke liye ek comprehension chahiye; DoL tumhe directly de deta hai. Woh saved gather exactly yahi wajah hai ki Pandas DataFrame (ek supercharged dict-of-lists) column-first build hoti hai.

Step 6 — Conversion LoD → DoL banao (code khud likhti hai)

KYA. Drawers ki shelf (Step 2) ko strips ke dict (Step 3) mein badlo, sirf wahi use karke jo pictures ne pehle se dikhaaya.

KYUN. Kyunki ab loop hi picture hai: "har column label ke liye, har row pe chalo aur us label ki value order mein collect karo" — yeh Step 3 ka vertical cut hai, zor se bola gaya.

PICTURE. Red mein highlighted: ek column assemble ho raha hai har drawer se neeche sweep karke.

Worked example LoD → DoL

lod = [{"name":"Asha","marks":88},
       {"name":"Ravi","marks":73},
       {"name":"Meera","marks":95}]
 
dol = {k: [row[k] for row in lod] for k in lod[0]}
# → {"name":["Asha","Ravi","Meera"], "marks":[88,73,95]}

for k in lod[0] — outer loop column labels par chalta hai, pehle drawer se udhar liya gaya (saare drawers labels share karte hain).
[row[k] for row in lod] — inner loop vertical cut hai: har drawer row ko sweep karo, row[k] nikalo, unhe row order mein rakho.

Step 7 — Reverse banao, DoL → LoD

KYA. Wapas rotate karo: dict of strips → shelf of drawers.

KYUN. Same idea, doosra axis. Ab hum ek row index i fix karte hain aur har column ki i-th cell padhte hain — yeh ek horizontal cut hai (Step 2), zor se bola gaya. Yeh literally students[c][i] ko wapas students[i][c] mein turn karna hai.

PICTURE. Red: ek row rebuild ho rahi hai har column strip ke position i mein pahunche ke.

Worked example DoL → LoD

dol  = {"name":["Asha","Ravi","Meera"], "marks":[88,73,95]}
keys = list(dol)                 # ["name","marks"]
n    = len(dol[keys[0]])         # rows = length of ANY column = 3
lod  = [{k: dol[k][i] for k in keys} for i in range(n)]
# → [{"name":"Asha","marks":88}, {"name":"Ravi","marks":73}, {"name":"Meera","marks":95}]

for i in range(n) — outer loop row positions $0, 1, 2$ par chalta hai.
{k: dol[k][i] for k in keys} — us fixed i ke liye, saare columns mein dol[k][i] padho → ek poora drawer. Yahan pair [k][i] exactly Step 3 ka [c][i] hai, Step 2 ki [i][c] row mein reassemble kiya gaya.

Step 8 — Degenerate cases (koi bhi aisa scenario mat aane do jo humne show nahi kiya)

KYA. Rotation ke sharp edges hain. Char jaanne wale hain, har ek ke saath picture ki kya toot ta hai.

KYUN. Contract: har case cover karo, empty aur ragged data sameti, taaki tum kabhi surprised na ho.

PICTURE. Left: ek empty grid. Middle: ek ragged grid (Ravi ke drawer mein marks missing hai). Right: "shared strip" trap jahan do column labels secretly ek hi physical list point karte hain.

Ek-picture summary

Upar sab kuch, compressed: same grid, do tarighon se cut kiya, red cell 73 ke saath jo left par [1]["marks"] se aur right par ["marks"][1] se milta hai — indices cross over karte hain, yahi transpose hai.

Recall Feynman retelling (plain words)

Ek class scoreboard ko table ki tarah draw karo. Agar tum use horizontal strips mein snip karo, har strip ek baache ki poori kahani hai — har strip ko ek labelled drawer mein rakho aur drawers ko ek shelf par khada karo: yeh ek list of dicts hai. Ravi ki marks paane ke liye tum kaho "drawer number 1, marks label kholo" → [1]["marks"]. Agar uske badle tum usi table ko vertical strips mein snip karo, har strip sabke liye ek tarah ka fact hai — har strip ko naam do aur unhe ek cupboard mein rakho: yeh ek dict of lists hai. Ab Ravi ki marks paane ke liye tum kaho "marks strip, line number 1" → ["marks"][1]. Same fact, lekin jo do words tum bolte ho unka order swap ho gaya hai — wahi swap poora secret hai, "transpose". Vertical strips se poore column ka sum karna trivial hai (woh pehle se ek line mein hain); horizontal drawers se ek baache ko grab karna ya add karna trivial hai (ek self-contained card). Unke beech convert karna bas table ko doosre tarike se re-snip karna hai — ek direction ke liye labels par loop, doosre ke liye positions par. Khaali tables, kisi label se khaali drawer, unequal length ki strips, aur wo sneaky bug se bachna jahan do labels bilkul same physical strip ko point karein.

Connections

Yeh note Hinglish mein → — same walkthrough, Hinglish
Lists — drawers ki shelf / column strips
Dictionaries — labelled drawers / strips ka cupboard
List Comprehensions — woh engine jo grid ko re-snip karta hai
Loops and Iteration — vertical aur horizontal sweeps
Mutable vs Immutable — kyun shared-strip bug (Case D) bites karta hai
Pandas DataFrame — column-first (DoL) idea, industrial strength
JSON and APIs — jahan list-of-dicts data usually aata hai