Git合并JSON文件时优化冲突对比可读性：实现块级而非片段式冲突展示

Absolutely! You can configure Git to handle JSON files in a structure-aware way, which will make merge conflicts show up at the object/array level instead of splitting into individual nested fields. Here's how to set it up:

1. Enable Block-Level JSON Diff & Conflicts #

Git's default diff treats JSON as plain text, leading to messy, fragmented conflicts when changes hit nested structures. By configuring Git to pretty-print JSON before diffing, you can make it recognize entire objects or array elements as single "blocks" of change.

Step 1: Install jq #

First, ensure you have jq (a lightweight JSON formatting tool) installed on your system—it’s used to format JSON into human-readable, line-separated structures.

Step 2: Configure Git for JSON Diff #

Add these settings to your global .gitconfig (or repository-specific .git/config):

[diff "json"]
    textconv = jq .
    cachetextconv = true

The textconv option tells Git to run jq . (which pretty-prints JSON with proper indentation) before generating diffs. cachetextconv speeds up repeated diffs by caching the formatted output.

Step 3: Link JSON Files to the Diff Driver #

Create or update your repository’s .gitattributes file to apply this driver to all JSON files:

*.json diff=json

Now, when you run git diff or encounter merge conflicts, Git will display JSON in a structured format. Conflicts will appear at the level of entire objects (like your desired example) instead of splitting into inner fields.

2. Automatic Merge for JSON Arrays (Optional) #

If you frequently add elements to JSON arrays (as in your scenario), you can set up a custom merge driver to automatically combine these additions instead of triggering a conflict.

Step 1: Create a Merge Script #

Make a script (e.g., json-merge-array.sh) with the following content, then mark it executable (chmod +x json-merge-array.sh):

#!/bin/bash
# Merge JSON arrays by combining unique elements from both branches
# Arguments: %O (base version), %A (local version), %B (remote version), %A (output file)
jq -s '.[0] + (.[1] - .[0]) + (.[2] - .[0])' "$1" "$2" "$3" > "$4"

This script uses jq to take the base JSON, add elements unique to your local branch, then add elements unique to the remote branch—perfect for cases where both branches add new array items.

Step 2: Configure the Merge Driver #

Add this to your .gitconfig:

[merge "json-array"]
    name = JSON Array Merge Driver
    driver = /full/path/to/json-merge-array.sh %O %A %B %A
    recursive = binary

Replace /full/path/to/ with the actual path to your script.

Step 3: Apply the Merge Driver #

Update your .gitattributes to use this driver for JSON files:

*.json merge=json-array diff=json

Now, when merging branches that both added elements to a JSON array, Git will automatically combine them into a single array with all items—no manual conflict resolution needed!

Note: If both branches modify the same JSON object (not just add new ones), you’ll still get a conflict—but it will be at the object level, making it far easier to resolve manually.

内容的提问来源于stack exchange，提问作者user15963491