python操作hbase(基于thrift服务)

一：环境说明

ubuntu 16.04
python2.7
happybase （python包, 执行命令 pip install happybase 安装）
hbase v1.3.1 (docker 镜像)
thrift

特别注意
thrift 、thrift2，新版本的hbase，默认使用thrift2，而thrift2相比thrift，去掉了很多对hbase的命令支持。如果你要换用thrift，只要停止thrift2 服务，启动thrift服务即可
启动、停止命令:
/hbase/bin/hbase-daemon.sh stop thrift2
/hbase/bin/hbase-daemon.sh start thrift

二：代码实现

# -*- coding:utf-8 -*-
import happybase
from collections import Iterable
import sys, os
HBASE_HOST = '127.0.0.1'
HBASE_PORT = 9090
class HBase(object):
    def __init__(self, _tableName, _families=None):
        self.connection = happybase.Connection(HBASE_HOST, HBASE_PORT, autoconnect=False)
        self.connection.open()
        self.tableName = _tableName
        self.create_table()

    def create_table(self):
        """
        如果此表不存在，则创建
        :param _tableName: 表名
        :return:
        """
        _tableList = self.connection.tables()
        families = {'%s' % self.tableName: dict()}
        if self.tableName not in _tableList:
            self.connection.create_table(self.tableName, families)

    def put(self, _rowKey, _fields):
        """
        insert/update
        :param _rowKey: string
        :param _fields: {'domain':'www.baidu.com', 'name':'百度'}
        :return:
        """
        print _rowKey, _fields
        fields = {}
        for k, v in _fields.items():
            fields['%s:%s' % (self.tableName, k)] = v

        table = self.connection.table(self.tableName)
        print _rowKey,fields
        table.put(_rowKey, fields)

    def scan(self, _rowKey=None, _filters=None):
        """
        多列、and、精确/模糊匹配 查询
        :param _rowKey: string  ^138$
        :param _filters: dict  {'name':'^百度$', 'domain':'^www.baidu.com$'}
        :return:
        """
        table = self.connection.table(self.tableName)
        filters = None
        if _rowKey:
            filters = "RowFilter(=,'regexstring:%s')" % _rowKey
        if _filters:
            for k, v in _filters.items():
                if not filters:
                    filters = "SingleColumnValueFilter('%s','%s',=,'regexstring:^%s$') " % (self.tableName, k, v)
                else:
                    fl = " AND SingleColumnValueFilter('%s','%s',=,'regexstring:^%s$')" % (self.tableName, k, v)
                    filters += fl

        rows = table.scan(filter=filters)
        return self.dumps(rows)

    def count(self, _field):
        """
        count统计行数
        :param _field: 根据某列统计
        :return: int
        """
        field = '%s:%s' % (self.tableName, _field)
        table = self.connection.table(self.tableName)
        count = 0
        for key, data in table.scan(columns=[field]):
            count += 1
        return count

    def row(self, _rowKey, _columns=None):
        """
        根据rowKey，获取某行数据
        :param _rowKey: string  '123'
        :param _columns: list ['name', 'domain']
        :return: dict
        """
        columns = []
        if _columns:
            columns = ['%s:%s' % (self.tableName, i) for i in _columns]

        table = self.connection.table(self.tableName)
        row = table.row(_rowKey, columns=columns)
        rtn = self.dumps(row)
        rtn['_id'] = _rowKey
        return rtn

    def rows(self, _rowKey):
        """
        根据rowKeys，获取某几行数据
        :param _rowKey: list ['123', '234']
        :return: list
        """
        table = self.connection.table(self.tableName)
        rows = table.rows(_rowKey)
        return self.dumps(rows)

    def dumps(self, _data):
        """
        将数据转换为dict，或list
        :param _data: list or dict or generator
        :return: dict or list
        """
        rtn = []
        if isinstance(_data, dict):
            rtn = {}
            for k, v in _data.items():
                rtn[k.split(':')[1]] = v
        elif isinstance(_data, list):
            for row in _data:
                tmp = {}
                tmp['_id'] = row[0]
                for k, v in row[1].items():
                    tmp[k.split(':')[1]] = v
                if tmp:
                    rtn.append(tmp)
        elif isinstance(_data, Iterable):
            for key, data in _data:
                tmp = {}
                tmp['_id'] = key
                for k, v in data.items():
                    tmp[k.split(':')[1]] = v
                if tmp:
                    rtn.append(tmp)
        else:
            pass

        return rtn

    def __del__(self):
        self.connection.close()


if __name__ =='__main__':
    print 'start----------------'
    hbase = HBase('test1')
    hbase.put('111', {'domain':'www.baidu.com', 'name':'百度'})
    print 'insert data finshed'
    print hbase.scan()
    print 'scan data finshed'
    print hbase.scan(_rowKey='1')
    print 'scan data finshed'
    print hbase.scan(_filters={'name':'^百度$', 'domain':'^www.baidu.com$'})
    print 'scan data finshed'
    print hbase.scan(_rowKey='^1$', _filters={'name': '^百度$', 'domain': '^www.baidu.com$'})
    print 'scan data finshed'
    print hbase.row('111')
    print 'row data finshed'
    print hbase.row('111', ['name'])
    print 'row data finshed'
    print hbase.rows(['111'])
    print 'rows data finshed'
    print hbase.count('name')
    print 'count data finshed'
    print 'end-----------------------'