12 KiB
NocoDB Automation Script Development Summary
Overview
This document summarizes the development of an automated NocoDB table creation script (build-nocodb.sh) for the Map Viewer project. The script automates the creation of three required tables: locations, login, and settings with proper schemas and default data.
Project Requirements
Based on the README.md analysis, the project needed:
- locations table: Main map data storage
- login table: User authentication
- settings table: System configuration and QR codes
- Default admin user and start location records
- Idempotent script (safe to re-run)
NocoDB API Research
API Versions
- v1 API:
/api/v1/- Legacy, limited functionality - v2 API:
/api/v2/- Modern, full-featured (recommended)
Key API Endpoints Discovered
Base/Project Management
GET /api/v2/meta/bases # List all bases
POST /api/v2/meta/bases # Create new base
GET /api/v2/meta/bases/{id} # Get base details
Table Management
GET /api/v2/meta/bases/{base_id}/tables # List tables in base
POST /api/v2/meta/bases/{base_id}/tables # Create table
GET /api/v2/meta/bases/{base_id}/tables/{table_id} # Get table details
Record Management
GET /api/v2/tables/{table_id}/records # List records
POST /api/v2/tables/{table_id}/records # Create record
PUT /api/v2/tables/{table_id}/records/{record_id} # Update record
Authentication
All API calls require the xc-token header:
curl -H "xc-token: YOUR_TOKEN" -H "Content-Type: application/json"
Table Schemas Implemented
1. Locations Table
Primary table for map data storage:
{
"table_name": "locations",
"columns": [
{"column_name": "id", "uidt": "ID", "pk": true, "ai": true},
{"column_name": "title", "uidt": "SingleLineText"},
{"column_name": "description", "uidt": "LongText"},
{"column_name": "category", "uidt": "SingleSelect", "colOptions": {
"options": [
{"title": "Important", "color": "#ff0000"},
{"title": "Event", "color": "#00ff00"},
{"title": "Business", "color": "#0000ff"},
{"title": "Other", "color": "#ffff00"}
]
}},
{"column_name": "geo_location", "uidt": "LongText"},
{"column_name": "latitude", "uidt": "Decimal"},
{"column_name": "longitude", "uidt": "Decimal"},
{"column_name": "address", "uidt": "LongText"},
{"column_name": "contact_info", "uidt": "LongText"},
{"column_name": "created_at", "uidt": "DateTime"},
{"column_name": "updated_at", "uidt": "DateTime"}
]
}
2. Login Table
User authentication table:
{
"table_name": "login",
"columns": [
{"column_name": "id", "uidt": "ID", "pk": true, "ai": true},
{"column_name": "username", "uidt": "SingleLineText", "rqd": true},
{"column_name": "email", "uidt": "Email", "rqd": true},
{"column_name": "password", "uidt": "SingleLineText", "rqd": true},
{"column_name": "admin", "uidt": "Checkbox"},
{"column_name": "active", "uidt": "Checkbox"},
{"column_name": "created_at", "uidt": "DateTime"},
{"column_name": "last_login", "uidt": "DateTime"}
]
}
3. Settings Table
System configuration with QR code support:
{
"table_name": "settings",
"columns": [
{"column_name": "id", "uidt": "ID", "pk": true, "ai": true},
{"column_name": "key", "uidt": "SingleLineText", "rqd": true},
{"column_name": "title", "uidt": "SingleLineText"},
{"column_name": "geo_location", "uidt": "LongText"},
{"column_name": "latitude", "uidt": "Decimal"},
{"column_name": "longitude", "uidt": "Decimal"},
{"column_name": "zoom", "uidt": "Number"},
{"column_name": "category", "uidt": "SingleSelect", "colOptions": {
"options": [
{"title": "system_setting", "color": "#4CAF50"},
{"title": "user_setting", "color": "#2196F3"},
{"title": "app_config", "color": "#FF9800"}
]
}},
{"column_name": "updated_by", "uidt": "SingleLineText"},
{"column_name": "updated_at", "uidt": "DateTime"},
{"column_name": "qr_code_1_url", "uidt": "URL"},
{"column_name": "qr_code_1_label", "uidt": "SingleLineText"},
{"column_name": "qr_code_1_image", "uidt": "Attachment"},
{"column_name": "qr_code_2_url", "uidt": "URL"},
{"column_name": "qr_code_2_label", "uidt": "SingleLineText"},
{"column_name": "qr_code_2_image", "uidt": "Attachment"},
{"column_name": "qr_code_3_url", "uidt": "URL"},
{"column_name": "qr_code_3_label", "uidt": "SingleLineText"},
{"column_name": "qr_code_3_image", "uidt": "Attachment"}
]
}
NocoDB Column Types (UIdt)
Discovered column types and their usage:
ID- Auto-incrementing primary keySingleLineText- Short text fieldLongText- Multi-line text areaEmail- Email validationURL- URL validationDecimal- Decimal numbersNumber- Integer numbersDateTime- Date and timeCheckbox- Boolean true/falseSingleSelect- Dropdown with predefined optionsAttachment- File upload field
Script Development Process
Initial Implementation
- Created basic structure with environment variable loading
- Implemented API connectivity testing
- Added base/project creation functionality
- Created table creation functions
Key Challenges Solved
1. Environment Variable Loading
Issue: Standard source .env wasn't exporting variables
Solution: Use set -a; source .env; set +a pattern
set -a # Auto-export all variables
source .env # Load environment file
set +a # Disable auto-export
2. API Version Compatibility
Issue: Mixed v1/v2 endpoint usage causing errors Solution: Standardized on v2 API with proper URL construction
BASE_URL=$(echo "$NOCODB_API_URL" | sed 's|/api/v1||')
API_BASE_V2="${BASE_URL}/api/v2"
3. Duplicate Table Error
Issue: Script failed when tables already existed Solution: Added idempotent table checking
get_table_id_by_name() {
local base_id=$1
local table_name=$2
# Check if table exists by name
local tables_response
tables_response=$(make_api_call "GET" "/meta/bases/$base_id/tables" "" "Fetching tables")
# Parse JSON to find table ID
local table_id
table_id=$(echo "$tables_response" | grep -o '"id":"[^"]*","table_name":"'"$table_name"'"' | grep -o '"id":"[^"]*"' | head -1 | sed 's/"id":"//;s/"//')
if [ -n "$table_id" ]; then
echo "$table_id"
return 0
else
return 1
fi
}
4. JSON Response Parsing
Issue: Complex JSON parsing for table IDs Solution: Used grep with regex patterns
# Extract table ID from JSON response
table_id=$(echo "$response" | grep -o '"id":"[^"]*"' | head -1 | cut -d'"' -f4)
Default Data Records
Admin User
{
"username": "admin",
"email": "admin@example.com",
"password": "changeme123",
"admin": true,
"active": true,
"created_at": "2025-07-05 12:00:00"
}
Start Location Setting
{
"key": "start_location",
"title": "Map Start Location",
"geo_location": "53.5461;-113.4938",
"latitude": 53.5461,
"longitude": -113.4938,
"zoom": 11,
"category": "system_setting",
"updated_by": "system",
"updated_at": "2025-07-05 12:00:00"
}
Error Handling Patterns
API Call Wrapper
make_api_call() {
local method=$1
local endpoint=$2
local data=$3
local description=$4
local api_version=${5:-"v2"}
# Construct full URL
if [[ "$api_version" == "v1" ]]; then
full_url="$API_BASE_V1$endpoint"
else
full_url="$API_BASE_V2$endpoint"
fi
# Make request with timeout
response=$(curl -s -w "%{http_code}" -X "$method" \
-H "xc-token: $NOCODB_API_TOKEN" \
-H "Content-Type: application/json" \
--max-time 30 \
-d "$data" \
"$full_url" 2>/dev/null)
# Parse HTTP code and response
http_code="${response: -3}"
response_body="${response%???}"
# Check for success
if [[ "$http_code" -ge 200 && "$http_code" -lt 300 ]]; then
echo "$response_body"
return 0
else
print_error "API call failed: $http_code - $response_body"
return 1
fi
}
Final Script Features
Idempotent Operation
- Checks for existing base/project
- Validates table existence before creation
- Uses existing table IDs when found
- Safe to run multiple times
Robust Error Handling
- Network timeout protection
- HTTP status code validation
- JSON parsing error handling
- Colored output for status messages
Environment Integration
- Loads configuration from
.envfile - Supports custom default coordinates
- Validates required variables
Usage Instructions
-
Setup Environment:
# Update .env with your NocoDB details NOCODB_API_URL=https://your-nocodb.com/api/v1 NOCODB_API_TOKEN=your_token_here -
Run Script:
chmod +x build-nocodb.sh ./build-nocodb.sh -
Post-Setup:
- Update
.envwith generated table URLs - Change default admin password
- Verify tables in NocoDB interface
- Update
Lessons Learned
- API Documentation: Always verify API endpoints with actual testing
- JSON Parsing: Shell-based JSON parsing requires careful regex patterns
- Idempotency: Essential for automation scripts in production
- Error Handling: Comprehensive error handling prevents silent failures
- Environment Variables: Proper loading patterns are crucial for script reliability
Future Enhancements
- Add support for custom table schemas via configuration
- Implement data migration features
- Add backup/restore functionality
- Support for multiple environment configurations
- Integration with CI/CD pipelines
Script Updates - July 2025
Column Type Improvements
Updated the build-nocodb.sh script to use proper NocoDB column types based on the official documentation:
Locations Table Updates
geo_location: Changed fromLongTexttoGeoData(proper geographic data type)latitude: Added precision (10) and scale (8) for proper decimal handlinglongitude: Added precision (11) and scale (8) for proper decimal handlingphone: Changed fromSingleLineTexttoPhoneNumber(proper phone validation)email: UsingEmailtype for proper email validation- Updated field names: Added proper fields from README.md:
first_name,last_name(SingleLineText)unit_number(SingleLineText)support_level(SingleSelect with colors: 1=Green, 2=Yellow, 3=Orange, 4=Red)sign(Checkbox)sign_size(SingleSelect: Small, Medium, Large)notes(LongText)address(SingleLineText instead of LongText)
Login Table Updates
- Simplified structure: Removed username/password fields per README.md specification
- Core fields:
email(Email),name(SingleLineText),admin(Checkbox) - Authentication note: This is a simplified table - proper authentication should be implemented separately
Settings Table Updates
geo_location: Changed fromLongTexttoGeoDatafor proper geographic data handlinglatitude/longitude: Added precision and scale parametersvalue: Added missingvaluefield from README.md specification- QR Code fields: Simplified to just attachment fields (removed URL/label fields not in README.md)
Benefits of Proper Column Types
-
GeoData Type:
- Proper latitude;longitude format validation
- Better integration with mapping libraries
- Consistent data storage format
-
PhoneNumber Type:
- Built-in phone number validation
- Proper formatting and display
- International number support
-
Email Type:
- Email format validation
- Prevents invalid email addresses
- Better UI experience
-
Decimal Precision:
- Latitude: 10 digits, 8 decimal places (±90.12345678)
- Longitude: 11 digits, 8 decimal places (±180.12345678)
- Provides GPS-level precision for mapping
-
SingleSelect with Colors:
- Support Level: Color-coded options for visual feedback
- Sign Size: Consistent option selection
- Category: Organized classification system
Backward Compatibility
The script maintains backward compatibility while using proper column types. Existing data migration may be needed if upgrading from the old schema.
Generated: July 5, 2025 Script Version: Column Type Optimized