[improvement](statistics)Skip auto analyze when table row count is not fully reported. (return -1)

fe/fe-core/src/main/java/org/apache/doris/catalog/OlapTable.java

@@ -138,6 +138,8 @@ public enum OlapTableState {
         WAITING_STABLE
     }
 
+    public static long ROW_COUNT_BEFORE_REPORT = -1;
+
     @SerializedName(value = "tst", alternate = {"state"})
     private volatile OlapTableState state;
 
@@ -1615,10 +1617,10 @@ public long getRowCountForIndex(long indexId, boolean strict) {
             if (index == null) {
                 LOG.warn("Index {} not exist in partition {}, table {}, {}",
                         indexId, entry.getValue().getName(), id, name);
-                return -1;
+                return ROW_COUNT_BEFORE_REPORT;
             }
             if (strict && !index.getRowCountReported()) {
-                return -1;
+                return ROW_COUNT_BEFORE_REPORT;
             }
             rowCount += index.getRowCount() == -1 ? 0 : index.getRowCount();
         }

fe/fe-core/src/main/java/org/apache/doris/statistics/StatisticsAutoCollector.java

@@ -149,6 +149,9 @@ protected void processOneJob(TableIf table, Set<Pair<String, String>> columns,
             return;
         }
         AnalysisInfo analyzeJob = createAnalyzeJobForTbl(table, columns, priority);
+        if (analyzeJob == null) {
+            return;
+        }
         LOG.debug("Auto analyze job : {}", analyzeJob.toString());
         try {
             executeSystemAnalysisJob(analyzeJob);
@@ -203,6 +206,13 @@ protected AnalysisInfo createAnalyzeJobForTbl(
         if (StatisticsUtil.enablePartitionAnalyze() && table.isPartitionedTable()) {
             analysisMethod = AnalysisMethod.FULL;
         }
+        if (table instanceof OlapTable && analysisMethod.equals(AnalysisMethod.SAMPLE)) {
+            OlapTable ot = (OlapTable) table;
+            if (ot.getRowCountForIndex(ot.getBaseIndexId(), true) == OlapTable.ROW_COUNT_BEFORE_REPORT) {
+                LOG.info("Table {} row count is not fully reported, skip auto analyzing this time.", ot.getName());
+                return null;
+            }
+        }
         AnalysisManager manager = Env.getServingEnv().getAnalysisManager();
         TableStatsMeta tableStatsStatus = manager.findTableStatsStatus(table.getId());
         long rowCount = StatisticsUtil.isEmptyTable(table, analysisMethod) ? 0 :

fe/fe-core/src/test/java/org/apache/doris/statistics/StatisticsAutoCollectorTest.java

@@ -138,4 +138,34 @@ public DLAType getDlaType() {
         ExternalTable hiveExternalTable = new HMSExternalTable(1, "hmsTable", "hmsDb", null);
         Assertions.assertTrue(collector.supportAutoAnalyze(hiveExternalTable));
     }
+
+    @Test
+    public void testCreateAnalyzeJobForTbl() {
+        StatisticsAutoCollector collector = new StatisticsAutoCollector();
+        OlapTable table = new OlapTable();
+        new MockUp<OlapTable>() {
+            @Mock
+            public long getDataSize(boolean singleReplica) {
+                return 100;
+            }
+
+            @Mock
+            public long getRowCountForIndex(long indexId, boolean strict) {
+                return -1;
+            }
+
+            @Mock
+            public boolean isPartitionedTable() {
+                return false;
+            }
+        };
+        Assertions.assertNull(collector.createAnalyzeJobForTbl(table, null, null));
+        new MockUp<OlapTable>() {
+            @Mock
+            public long getRowCountForIndex(long indexId, boolean strict) {
+                return 100;
+            }
+        };
+        Assertions.assertThrows(NullPointerException.class, () -> collector.createAnalyzeJobForTbl(table, null, null));
+    }
 }

regression-test/suites/statistics/test_auto_analyze_black_white_list.groovy

@@ -17,6 +17,36 @@
 
 suite("test_auto_analyze_black_white_list") {
 
+    def wait_row_count_reported = { db, table, row, column, expected ->
+        def result = sql """show frontends;"""
+        logger.info("show frontends result origin: " + result)
+        def host
+        def port
+        for (int i = 0; i < result.size(); i++) {
+            if (result[i][8] == "true") {
+                host = result[i][1]
+                port = result[i][4]
+            }
+        }
+        def tokens = context.config.jdbcUrl.split('/')
+        def url=tokens[0] + "//" + host + ":" + port
+        logger.info("Master url is " + url)
+        connect(user = context.config.jdbcUser, password = context.config.jdbcPassword, url) {
+            sql """use ${db}"""
+            result = sql """show frontends;"""
+            logger.info("show frontends result master: " + result)
+            for (int i = 0; i < 120; i++) {
+                Thread.sleep(5000)
+                result = sql """show index stats ${table} ${table};"""
+                logger.info("result " + result)
+                if (result[row][column] == expected) {
+                    return;
+                }
+            }
+            throw new Exception("Row count report timeout.")
+        }
+    }
+
     sql """drop database if exists test_auto_analyze_black_white_list"""
     sql """create database test_auto_analyze_black_white_list"""
     sql """use test_auto_analyze_black_white_list"""
@@ -38,6 +68,13 @@ suite("test_auto_analyze_black_white_list") {
         )
     """
 
+    try {
+        wait_row_count_reported("test_auto_analyze_black_white_list", "test_bw", 0, 4, "0")
+    } catch (Exception e) {
+        logger.info(e.getMessage());
+        return;
+    }
+
     // Test show index row count
     def result = sql """show table stats test_bw"""
     assertEquals(1, result.size())