oracle查所有表的索引个数的示例代码实用指南

作者：袖梨 2026-09-01

平时做技术实践时，很多问题不是概念不会，而是细节没串起来。拿“oracle查所有表的索引个数的示例代码”来说，它看着像小点，放到项目里常会牵出环境、配置、兼容性和维护成本。下面按实际采用顺序，把思路、关键写法和容易踩坑的地方讲清楚，便于大家直接对照操作。

1. 查看当前用户所有表的索引数量

SELECT 
    t.table_name,
    COUNT(i.index_name) as index_count,
    LISTAGG(i.index_name, ', ') WITHIN GROUP (ORDER BY i.index_name) as index_names
FROM user_tables t
LEFT JOIN user_indexes i ON t.table_name = i.table_name
GROUP BY t.table_name
ORDER BY COUNT(i.index_name) DESC, t.table_name;

2. 查看所有用户所有表的索引数量（需DBA权限）

SELECT 
    i.table_owner,
    i.table_name,
    COUNT(i.index_name) as index_count,
    LISTAGG(i.index_name, ', ') WITHIN GROUP (ORDER BY i.index_name) as index_names
FROM dba_indexes i
WHERE i.table_owner NOT IN ('SYS', 'SYSTEM', 'XDB', 'CTXSYS', 'MDSYS', 'ORDSYS')  -- 排除系统用户
GROUP BY i.table_owner, i.table_name
ORDER BY i.table_owner, COUNT(i.index_name) DESC, i.table_name;

3.查询表及其索引的详细信息落到代码里，—建议采用，oracle国产化转换到tidb，最好明确知道所有需迁移的生产表的条数等令牌

SELECT 
    t.owner,
    t.table_name,
    t.num_rows as table_rows,
    COUNT(i.index_name) as total_indexes,
    SUM(CASE WHEN i.uniqueness = 'UNIQUE' THEN 1 ELSE 0 END) as unique_indexes,
    SUM(CASE WHEN i.uniqueness = 'NONUNIQUE' THEN 1 ELSE 0 END) as nonunique_indexes,
    SUM(CASE WHEN i.index_type = 'FUNCTION-BASED NORMAL' THEN 1 ELSE 0 END) as function_based_indexes
FROM dba_tables t
LEFT JOIN dba_indexes i ON t.owner = i.table_owner AND t.table_name = i.table_name
WHERE t.owner = 'YOUR_SCHEMA_NAME'  -- 替换为你的模式名
GROUP BY t.owner, t.table_name, t.num_rows
ORDER BY COUNT(i.index_name) DESC, t.table_name;

4.按索引类型统计

SELECT 
    i.table_owner,
    i.table_name,
    i.index_type,
    COUNT(*) as count_per_type,
    LISTAGG(i.index_name, ', ') WITHIN GROUP (ORDER BY i.index_name) as index_list
FROM dba_indexes i
WHERE i.table_owner = 'YOUR_SCHEMA_NAME'  -- 替换为你的模式名
GROUP BY i.table_owner, i.table_name, i.index_type
ORDER BY i.table_name, i.index_type;

5.查询没有索引的表

-- 查找当前用户下没有索引的表
SELECT 
    t.table_name,
    t.num_rows,
    t.blocks
FROM user_tables t
WHERE NOT EXISTS (
    SELECT 1 
    FROM user_indexes i 
    WHERE i.table_name = t.table_name
)
AND t.table_name NOT LIKE 'BIN$%'  -- 排除回收站中的表
ORDER BY t.num_rows DESC NULLS LAST;

-- 查找所有用户下没有索引的表（需要DBA权限）
SELECT 
    t.owner,
    t.table_name,
    t.num_rows
FROM dba_tables t
WHERE NOT EXISTS (
    SELECT 1 
    FROM dba_indexes i 
    WHERE i.table_owner = t.owner 
    AND i.table_name = t.table_name
)
AND t.owner NOT IN ('SYS', 'SYSTEM', 'XDB')
AND t.table_name NOT LIKE 'BIN$%'
ORDER BY t.owner, t.num_rows DESC NULLS LAST;

6.索引列数统计

-- 统计每个索引的列数
SELECT 
    i.table_name,
    i.index_name,
    i.uniqueness,
    i.status,
    COUNT(ic.column_position) as column_count,
    LISTAGG(ic.column_name, ', ') WITHIN GROUP (ORDER BY ic.column_position) as columns
FROM user_indexes i
JOIN user_ind_columns ic ON i.index_name = ic.index_name
GROUP BY i.table_name, i.index_name, i.uniqueness, i.status
ORDER BY i.table_name, i.index_name;

7.实用的汇总查询

-- 索引统计汇总
WITH index_stats AS (
    SELECT 
        owner,
        table_name,
        COUNT(*) as total_indexes,
        ROUND(AVG(blevel), 2) as avg_blevel,
        ROUND(AVG(leaf_blocks), 2) as avg_leaf_blocks,
        SUM(CASE WHEN status != 'VALID' THEN 1 ELSE 0 END) as invalid_indexes
    FROM dba_indexes
    WHERE owner = 'YOUR_SCHEMA_NAME'
    GROUP BY owner, table_name
)
SELECT 
    owner,
    COUNT(DISTINCT table_name) as tables_with_indexes,
    SUM(total_indexes) as total_index_count,
    ROUND(AVG(total_indexes), 2) as avg_indexes_per_table,
    ROUND(MEDIAN(total_indexes), 2) as median_indexes_per_table,
    MAX(total_indexes) as max_indexes_in_table,
    SUM(invalid_indexes) as total_invalid_indexes
FROM index_stats
GROUP BY owner;

8.生产监控大表无索引情况

-- 查找行数超过10000但索引数少于2个的表
SELECT 
    t.owner,
    t.table_name,
    t.num_rows,
    COUNT(i.index_name) as index_count
FROM dba_tables t
LEFT JOIN dba_indexes i ON t.owner = i.table_owner AND t.table_name = i.table_name
WHERE t.num_rows > 10000
AND t.owner = 'YOUR_SCHEMA_NAME'
GROUP BY t.owner, t.table_name, t.num_rows
HAVING COUNT(i.index_name) < 2
ORDER BY t.num_rows DESC;

9.查看索引采用情况（需Oracle 11g及以上）

SELECT 
    table_name,
    index_name,
    used
FROM v$object_usage
WHERE used = 'NO'  -- 查看未使用的索引
ORDER BY table_name;

10.生成新建索引的脚本

SELECT 
    'CREATE INDEX idx_' || table_name || '_' || column_name || 
    ' ON ' || table_name || '(' || column_name || ');' as create_index_sql
FROM (
    SELECT DISTINCT
        t.table_name,
        tc.column_name
    FROM user_tables t
    JOIN user_tab_columns tc ON t.table_name = tc.table_name
    WHERE NOT EXISTS (
        SELECT 1 
        FROM user_ind_columns ic 
        WHERE ic.table_name = t.table_name 
        AND ic.column_name = tc.column_name
    )
    AND t.table_name NOT LIKE 'BIN$%'
    AND tc.column_name NOT LIKE '%ID'  -- 排除ID列
    AND tc.data_type IN ('VARCHAR2', 'CHAR', 'NUMBER', 'DATE')  -- 只对某些数据类型创建索引
)
WHERE ROWNUM <= 10;  -- 限制生成的数量