统计聚合

参数

示例

统计年龄为 18 岁的人中得分的最低分数。

query = TermQuery('age', 18)
agg = Min('score', name='min')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

统计年龄为 18 岁的人中得分的最高分数。如果某人没有分数，则对应分数的默认值为 0。

query = TermQuery('age', 18)
agg = Max('score', missing_value=0, name='max')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

统计年龄为 18 岁的所有人得分的总和。

query = TermQuery('age', 18)
agg = Sum('score', name='sum')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

统计年龄为 18 岁的所有人得分的平均分。

query = TermQuery('age', 18)
agg = Avg('score', name='avg')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

统计年龄为 18 岁的人中参加考试有分数的人数。

query = TermQuery('age', 18)
agg = Count('score', name='count')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

去重统计年龄为 18 岁的人中一共有多少种不同的姓名。

query = TermQuery('age', 18)
agg = DistinctCount('name', name='distinct_name')

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s, value: %s" % (agg_result.name, str(agg_result.value)))

参数

示例

query = TermQuery('product', '10010')
agg = Percentiles('latency', percentiles_list=[50, 90, 95])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, aggs=[agg]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for agg_result in search_response.agg_results:
    print("name: %s" % agg_result.name)
    for item in agg_result.value:
        print("  percentile: %s, value: %s" % (str(item.key), str(item.value)))

参数

示例 1

将年龄为 18 岁的人按分数分组，并获取人数最多的 10 个分数值以及每个分数的人数。

query = TermQuery('age', 18)
group_by = GroupByField('score', size=10)

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  key: %s, count: %d" % (item.key, item.row_count))

示例 2

将年龄为 18 岁的人按分数分组，并获取人数最少的 2 个分数值以及每个分数的人数。

group_by = GroupByField('score', size=2, group_by_sort=[RowCountSort(sort_order=SortOrder.ASC)])

search_response = client.search(table_name, index_name,
    SearchQuery(TermQuery('age', 18), limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  key: %s, count: %d" % (item.key, item.row_count))

示例 3

将年龄为 18 岁的人按分数分组，并获取人数最多的 2 个分数值、每个分数的人数以及按主键排序前三的人的信息。

sort = RowCountSort(sort_order=SortOrder.DESC)
sub_agg = [TopRows(limit=3, sort=Sort([PrimaryKeySort(sort_order=SortOrder.DESC)]), name='top_rows')]
group_by = GroupByField('score', size=2, group_by_sort=[sort], sub_aggs=sub_agg)

search_response = client.search(table_name, index_name,
    SearchQuery(TermQuery('age', 18), limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  key: %s, count: %d" % (item.key, item.row_count))
        for sub_agg in item.sub_aggs:
            print("    sub_agg: %s" % sub_agg.name)
            for entry in sub_agg.value:
                print("      value: %s" % str(entry))

示例 4

将年龄为 18 岁的人按分数和性别分组。

sort = RowCountSort(sort_order=SortOrder.ASC)
sub_group = GroupByField('sex', size=10, group_by_sort=[sort])
group_by = GroupByField('score', size=10, group_by_sort=[sort], sub_group_bys=[sub_group])

search_response = client.search(table_name, index_name,
    SearchQuery(TermQuery('age', 18), limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  key: %s, count: %d" % (item.key, item.row_count))
        for sub_group in item.sub_group_bys:
            print("    sub_group: %s" % sub_group.name)
            for sub_item in sub_group.items:
                print("      key: %s, count: %s" % (str(sub_item.key), str(sub_item.row_count)))

参数

示例

统计年龄为 18 岁的人中得分的分数在 [80, 90) 和 [90, 100) 两个区间段的人数。

query = TermQuery('age', 18)
group_by = GroupByRange(field_name='score', ranges=[(80, 90), (90, 100)])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  range: %.1f~%.1f, count: %d" % (item.range_from, item.range_to, item.row_count))

参数

示例

统计年龄为 18 岁的人中家庭住址在距离学校一公里以内和一公里到两公里内的人数。其中学校经纬度为(31,116)。

query = TermQuery('age', 18)
group_by = GroupByGeoDistance(field_name='address', origin=GeoPoint(31, 116), ranges=[(0, 1000), (1000, 2000)])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  range: %.1f~%.1f, count: %d" % (item.range_from, item.range_to, item.row_count))

参数

示例

分别统计年龄为 18 岁的人中数学考了 100 分和语文考了 100 分的人数。

query = TermQuery('age', 18)
filter1 = TermQuery('math', 100)
filter2 = TermQuery('chinese', 100)
filters = [filter1, filter2]
group_by = GroupByFilter(filters)

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    i = 0
    for item in group_by_result.items:
        print("  filter: %s=%s, count: %d" % (str(filters[i].field_name), str(filters[i].column_value), item.row_count))
        i += 1

参数

示例

query = TermQuery('product', '10010')
group_by = GroupByHistogram(field_name='latency', interval=100, field_range=FieldRange(0, 10000), missing_value=0)

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.ALL_FROM_INDEX))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  key: %s, value: %s" % (item.key, item.value))

参数

示例 1

按字段 score 对年龄为 18 岁的人进行分组，获取每个分数值对应的行数。

query = TermQuery('age', 18)
# 作为数据源的 GroupByField 只需指定 field_name 和 name，不能设置 size
source = GroupByField('score', name='group_by_score')
group_by = GroupByComposite(sources=[source])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.NONE))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    print("source_names: %s" % str(group_by_result.source_group_by_names))
    for item in group_by_result.items:
        print("  keys: %s, count: %d" % (str(item.keys), item.row_count))

示例 2

按字段 score 和 sex 对年龄为 18 岁的人进行组合分组，获取每个分数和性别组合对应的行数。

query = TermQuery('age', 18)
source1 = GroupByField('score', name='group_by_score')
source2 = GroupByField('sex', name='group_by_sex')
group_by = GroupByComposite(sources=[source1, source2])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.NONE))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    print("source_names: %s" % str(group_by_result.source_group_by_names))
    for item in group_by_result.items:
        # keys 是字符串列表，与 sources 顺序一一对应；字段值为空时对应位置为 None
        print("  keys: %s, count: %d" % (str(item.keys), item.row_count))

示例 3

使用 next_token 实现分页获取所有分组结果。按字段 score 分组，每次获取 2 个分组，通过翻页获取所有结果。

query = TermQuery('age', 18)
source = GroupByField('score')
group_by = GroupByComposite(sources=[source], size=2)

# 首次请求
search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.NONE))

group_by_result = search_response.group_by_results[0]
all_items = list(group_by_result.items)

# 翻页获取剩余结果，next_token 为 None 时表示已获取完毕
while group_by_result.next_token is not None:
    group_by = GroupByComposite(sources=[source], size=2, next_token=group_by_result.next_token)
    search_response = client.search(table_name, index_name,
        SearchQuery(query, limit=0, get_total_count=True, group_bys=[group_by]),
        ColumnsToGet(return_type=ColumnReturnType.NONE))

    group_by_result = search_response.group_by_results[0]
    all_items.extend(group_by_result.items)

# 输出所有分组结果
for item in all_items:
    print("keys: %s, count: %d" % (str(item.keys), item.row_count))

示例 4

使用子统计聚合。按字段 score 分组后，统计每个分组中 score 的最大值。

query = TermQuery('age', 18)
source = GroupByField('score')
sub_agg = Max('score')
group_by = GroupByComposite(sources=[source], sub_aggs=[sub_agg])

search_response = client.search(table_name, index_name,
    SearchQuery(query, limit=0, get_total_count=True, group_bys=[group_by]),
    ColumnsToGet(return_type=ColumnReturnType.NONE))

for group_by_result in search_response.group_by_results:
    print("name: %s" % group_by_result.name)
    for item in group_by_result.items:
        print("  keys: %s, count: %d" % (str(item.keys), item.row_count))
        for sub_agg in item.sub_aggs:
            print("    sub_agg: %s, value: %s" % (sub_agg.name, str(sub_agg.value)))