Dates & Times

Every price, every trade, every candle carries a timestamp - and timestamps are where tidy data gets messy. Different formats, time zones, weekends, market hours, holidays: dates and times have more quirks than any other kind of data. The good news is Python's datetime module handles the hard parts, and once you know a few moves, working with dates becomes routine. This chapter gives the timestamp the respect it deserves.

date and datetime

There are two everyday types. A date is a calendar day (no time); a datetime is a day and a time. You build them by passing the parts, and read the parts back as attributes:

from datetime import date, datetime

d = date(2026, 6, 25)                     # a calendar date
dt = datetime(2026, 6, 25, 15, 30, 0)     # a date AND a time (3:30 PM, market close)

print("Date     :", d)
print("Datetime :", dt)
print("Parts    : year", d.year, "month", d.month, "day", d.day)
print("Time     : hour", dt.hour, "minute", dt.minute)
print("Weekday  :", d.strftime("%A"))     # the day name

Date     : 2026-06-25
Datetime : 2026-06-25 15:30:00
Parts    : year 2026 month 6 day 25
Time     : hour 15 minute 30
Weekday  : Thursday

date(2026, 6, 25) is the 25th of June; add 15, 30 and you've got datetime(2026, 6, 25, 15, 30) - half past three in the afternoon, the moment the NSE cash market closes. Every part - .year, .month, .day, .hour, .minute - is right there to read.

Formatting and parsing: strftime and strptime

Timestamps constantly need to move between datetime objects (which you can do maths on) and text (which humans and files use). Two methods, with a handy mnemonic, do the conversion:

from datetime import datetime

dt = datetime(2026, 6, 25, 15, 30)

# strftime: a datetime -> formatted TEXT   (f = format)
print("ISO date  :", dt.strftime("%Y-%m-%d"))
print("Readable  :", dt.strftime("%d %b %Y, %I:%M %p"))
print("Day & time:", dt.strftime("%A %H:%M"))

# strptime: TEXT -> a datetime   (p = parse)
text = "2026-06-25 09:15"
parsed = datetime.strptime(text, "%Y-%m-%d %H:%M")
print("Parsed    :", parsed, "| type:", type(parsed).__name__)

ISO date  : 2026-06-25
Readable  : 25 Jun 2026, 03:30 PM
Day & time: Thursday 15:30
Parsed    : 2026-06-25 09:15:00 | type: datetime

Both use the same little format codes - %Y for year, %H for hour, and so on. Here's a cheat-sheet of the ones you'll reach for most:

Date arithmetic with timedelta

To add or subtract time, use a timedelta - a duration. Adding one to a datetime gives you another datetime:

from datetime import datetime, timedelta, time

# timedelta does date and time arithmetic.
close = datetime(2026, 6, 25, 15, 30)
print("Close         :", close)
print("One week later:", close + timedelta(weeks=1))
print("90 min before :", close - timedelta(minutes=90))

# Is a given moment within NSE cash hours (09:15 to 15:30, weekdays only)?
def is_market_open(dt):
    return dt.weekday() < 5 and time(9, 15) <= dt.time() <= time(15, 30)

print("11:00 Thu open?", is_market_open(datetime(2026, 6, 25, 11, 0)))   # True
print("16:00 Thu open?", is_market_open(datetime(2026, 6, 25, 16, 0)))   # False (after close)
print("11:00 Sun open?", is_market_open(datetime(2026, 6, 28, 11, 0)))   # False (weekend)

Close         : 2026-06-25 15:30:00
One week later: 2026-07-02 15:30:00
90 min before : 2026-06-25 14:00:00
11:00 Thu open? True
16:00 Thu open? False
11:00 Sun open? False

timedelta(weeks=1) jumps a week ahead; timedelta(minutes=90) steps back 90 minutes. And the is_market_open function shows the real payoff: combining .weekday() (where Monday is 0 and Saturday/Sunday are 5 and 6) with a time range to ask "is this moment inside NSE trading hours, on a weekday?" That single idea - filtering data to market hours, excluding weekends - is something you'll do constantly with real price data.

Try it yourself

• Format datetime(2026, 6, 25, 9, 15) as "09:15 AM on Thursday" using the right % codes.
• Parse the text "25-12-2026" into a date with strptime and the format "%d-%m-%Y". What weekday is Christmas 2026?
• Use timedelta(days=...) to find the date 45 days after date(2026, 6, 25) - a rough options-expiry horizon.

Recap

• date is a calendar day; datetime adds a time. Read parts via .year, .month, .hour, etc.
• strftime formats a datetime to text, strptime parses text to a datetime - both use % codes (f = format out, p = parse in).
• timedelta represents a duration; add or subtract it to do date arithmetic.
• .weekday() (Mon = 0) plus a time range filters data to weekday market hours - a constant need with price data.

That's nearly all of Module 3. You can import libraries, install packages, read and write files and JSON, handle errors, and work with dates. The final piece is a meta-skill - the one that makes you self-sufficient for everything still ahead: knowing how to get unstuck on your own. It's the next, and last, chapter of Module 3.