CSV and Text

CSV to Dicts

Convert parsed CSV rows into a list of dicts by zipping the header row with each data row. The trace shows rec being built fresh for each row via dict(zip(header, fields)) before being appended to records.

By hand

Split lines as in parse-csv-rows, treat the first line as the header, then for each data line build a dict with dict(zip(header, line.split(','))). records is ignored in the trace so the per-row rec dict is visible.

naive.py
Replay: real traced execution (multi-file project)
csv_text = 'id,cat,val\n1,x,10\n2,y,20\n3,x,30'
lines = csv_text.split('\n')
header = lines[0].split(',')
records = []
# trace: ignore records
for line in lines[1:]:
    rec = dict(zip(header, line.split(',')))
    records.append(rec)
print('RESULT:', records[-1])

  1. csv_text ← 'id,cat,val\n1,x,10\n2,y,20\n3,x,30'

    1csv_text = 'id,cat,val\n1,x,10\n2,y,20\n3,x,30'2lines = csv_text.split('\n')
    values this step'id,cat,val\n1,x,10\n2,y,20\n3,x,30'csv_text

  2. lines ← ['id,cat,val', '1,x,10', '2,y,20', '3,x,30']

    1csv_text = 'id,cat,val\n1,x,10\n2,y,20\n3,x,30'2lines = csv_text.split('\n')3header = lines[0].split(',')
    values this step['id,cat,val', '1,x,10', '2,y,20', '3,x,30']lines

  3. header ← ['id', 'cat', 'val']

    2lines = csv_text.split('\n')3header = lines[0].split(',')4records = []
    values this step['id', 'cat', 'val']header

  4. records = []

    3header = lines[0].split(',')4records = []5# trace: ignore records

  5. line ← '1,x,10'

    5# trace: ignore records6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))
    values this step'1,x,10'line

  6. rec ← {'id': '1', 'cat': 'x', 'val': '10'}

    6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))8    records.append(rec)
    values this step{'id': '1', 'cat': 'x', 'val': '10'}rec

  7. records.append(rec)

    7    rec = dict(zip(header, line.split(',')))8    records.append(rec)9print('RESULT:', records[-1])

  8. line ← '2,y,20'

    5# trace: ignore records6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))
    values this step'1,x,10' → '2,y,20'line

  9. rec ← {'id': '2', 'cat': 'y', 'val': '20'}

    6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))8    records.append(rec)
    values this step{'id': '1', 'cat': 'x', 'val': '10'} → {'id': '2', 'cat': 'y', 'val': '20'}rec

  10. records.append(rec)

    7    rec = dict(zip(header, line.split(',')))8    records.append(rec)9print('RESULT:', records[-1])

  11. line ← '3,x,30'

    5# trace: ignore records6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))
    values this step'2,y,20' → '3,x,30'line

  12. rec ← {'id': '3', 'cat': 'x', 'val': '30'}

    6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))8    records.append(rec)
    values this step{'id': '2', 'cat': 'y', 'val': '20'} → {'id': '3', 'cat': 'x', 'val': '30'}rec

  13. records.append(rec)

    7    rec = dict(zip(header, line.split(',')))8    records.append(rec)9print('RESULT:', records[-1])

  14. for line in lines[1:]:

    5# trace: ignore records6for line in lines[1:]:7    rec = dict(zip(header, line.split(',')))

  15. stdout ← RESULT: {'id': '3', 'cat': 'x', 'val': '30'}

    8    records.append(rec)9print('RESULT:', records[-1])
    values this stepRESULT: {'id': '3', 'cat': 'x', 'val': '30'}stdout

The Pythonic way

csv.DictReader reads the header automatically and yields one dict per data row. list(...) materialises the result.

library.py 
import csv
import io
csv_text = 'id,cat,val\n1,x,10\n2,y,20\n3,x,30'
records = list(csv.DictReader(io.StringIO(csv_text)))
print('RESULT:', records[-1])

RESULT: {'id': '3', 'cat': 'x', 'val': '30'}

Implementation notes

  • RESULT prints the last dict; the full list of 3 dicts would exceed the 80-char repr limit.
  • dict(zip(header, fields)) is the building block behind csv.DictReader — both pair column names with values positionally.
  • All field values are strings after CSV parsing; use int(rec['val']) or similar to convert numeric columns.
  • See parse-csv-rows (this chapter) for the prior step that produces the raw row lists this lesson consumes.