cnloc (China Location)

Last updated on 2026-01-13

cnloc is an Stata interface for the Python cnloc package, which parses Chinese addresses to get province, city, and county names (_name), administrative division codes (_adcode), and unique IDs (_id) from addresses.

This command requires Python package cnloc.

pip install cnloc

Usage

Installation:

net install cnloc, from("https://raw.githubusercontent.com/maobin-xu/cnloc-stata/main/") replace

Command syntax:

cnloc address_var [ year_var] [if] [in] [, year(str) mode(str) drop(str) prefix(str) suffix(str) county_short ]

• address_var: variable name for address strings
• year_var: variable name for year
• year(str): year for matching. An interger or variable name. Default is year(2024). Valid range is from 1980 to 2024, and invalid years will be clipped to the default year 2024. Alternative method is to use year_var
• mode(str): matching mode. Default is mode(1). mode(1) is left to right (high to low, province to county), following string order. mode(2) is low to high (county to province), ignoring string order; not recommended for basic use
• drop(str): variable list to drop in the final output. Default is None. drop(year) drops the year variable. drop(name)} drops full-name variables of province_name, city_name, and county_name. drop(adcode) drops administrative-division-code variables of province_adcode, city_adcode, and county_adcode. drop(id) drops unique IDs of province_id, city_id, and county_id variables
• prefix(str): prefix to add to variable names
• suffix(str): suffix to add to variable names
• county_short: allow short names for county-level matching. Default is not.

Currently, the county_id is unreliable! Only province- and city-level IDs are manually verified.

Examples

Prepare sample address data

clear
set obs 34
gen address1 = ""
local address_list `" "朝阳" "朝阳市" "朝阳县" "朝阳区" "朝阳市朝阳" "朝阳朝阳" "北京朝阳" "辽宁朝阳" "荆州" "荆州市" "荆州区" "荆州市荆州区" "荆州市荆州" "荆州荆州" "荆州荆州区" "湖北荆州" "湖北荆州沙市" "鼓楼区" "江苏鼓楼区" "南京鼓楼区" "江苏徐州鼓楼区" "南山" "广东省深圳市南山区深南大道" "深圳南山" "广东南山" "深圳市华侨城东部工业区" "深圳东门南路2006号宝丰大厦五楼" "中国深圳市深南大道" "海淀" "北京市海淀区中关村大街1号" "海淀中关村大街1号" "中关村大街1号" "马鞍山市经济技术开发区红旗南路51号" "银川市西夏区北京西路630号" "'

* address
local i = 1
foreach addr in `address_list' {
    replace address1 = "`addr'" if _n == `i'
    local i = `i' + 1
}

* parse year
gen year1 = 2024

Example 1: A simple example. The two are identical.

cnloc address1
cnloc address1 , year(2024)

Example 2: Use user-specified parse year.

cnloc address1 year1

Example 3: Use parse year of 2022. Drop parse year, names and administrative division codes of provinces, cities, and counties. Set prefix and suffix for final variables. Allow matching using county short names.

cnloc address1 , year(2022) drop(year name adcode) prefix(a_) suffix(_b) county_short