Calculate Average PR Merge Time

Calculate Average PR Merge Time

Overview

Calculate the average time-to-merge in days (rounded to 1 decimal place) for PRs that were successfully merged during December 2024.

Step 1: Filter Merged PRs Only

From the PR query results, identify only PRs where:

• mergedAt is NOT null
• state equals "MERGED" (verify this is the actual state value)

Do NOT include closed-but-unmerged or open PRs in this calculation.

Step 2: Parse Timestamps Correctly

For each merged PR, extract:

• createdAt: The PR creation timestamp (ISO-8601 format, e.g., 2024-12-15T10:30:45Z)
• mergedAt: The PR merge timestamp (ISO-8601 format)

Critical: Use the creation date as the start point, not any other date field.

Step 3: Convert to Days

For each PR:

1. Parse both timestamps as datetime objects (handle timezone info — typically UTC)
2. Calculate the difference: mergedAt - createdAt
3. Convert to total days (may be a float with decimal)
4. Example: If a PR was created on 2024-12-01 at 00:00:00 and merged on 2024-12-02 at 12:00:00, that's 1.5 days

Step 4: Compute Average

• Sum all day values from step 3
• Divide by the count of merged PRs
• Round the result to exactly 1 decimal place (e.g., 1.2, 15.8, 0.5)

Step 5: Validation Check

If your average is significantly higher than expected (e.g., 16+ days):

• Verify you're only including merged PRs (check mergedAt != null)
• Confirm you're measuring from createdAt (not some other field)
• Check if any timestamp parsing errors are inflating the values
• Inspect a few sample PRs manually to spot-check the calculation

Example Pseudocode

merged_prs = [pr for pr in prs if pr.mergedAt is not null and pr.state == "MERGED"]
days_list = []
for pr in merged_prs:
  created = parse_datetime(pr.createdAt)
  merged = parse_datetime(pr.mergedAt)
  days = (merged - created).total_seconds() / 86400
  days_list.append(days)
avg = sum(days_list) / len(days_list)
rounded_avg = round(avg, 1)