您好,并提前致谢。我使用RODBC
与Netezza数据库建立ODBC连接,我正在尝试将数据帧写入数据库表。我一直收到以下错误:
HY000 46 ERROR: Numeric overflow. Value does not fit in a numeric(10,10)
我不明白我是如何得到它的,因为我的所有数值都四舍五入到最接近的数字!我有我的SQL查询,我在下面创建了表和我的R代码,并提供了我正在尝试写入数据库的示例:
CREATE TABLE R_FORECAST
(
ID INTEGER NOT NULL,
NUMBER CHARACTER(7) NOT NULL,
ITEM_NO INTEGER NOT NULL,
ITEM_DESC VARCHAR(50),
FORECASTDATE DATE NOT NULL,
MEANFORECAST NUMERIC(10,10),
UPPER50 NUMERIC(10,10),
UPPER75 NUMERIC(10,10),
UPPER90 NUMERIC(10,10),
LOWER50 NUMERIC(10,10),
LOWER75 NUMERIC(10,10),
LOWER90 NUMERIC(10,10),
CREATED_DT TIMESTAMP DEFAULT "TIMESTAMP"(now(0)::"VARCHAR"),
TRANSACTION_DT DATE NOT NULL
)
DISTRIBUTE ON (ID);
loadprod <- odbcConnect(dsn="dsn", uid = "uid", pwd = "pwd",case="nochange", colQuote=c() )
head(FORECAST)
TRANSACTION_DT ID NUMBER ITEM_NO ITEM_DESC FORECASTDATE MEANFORECAST UPPER50 UPPER75 UPPER90 LOWER50 LOWER75 LOWER90 CREATED_DT
1/26/2015 4192 2239 9664 DESC 2/16/2015 11 13 17 22 9 7 2 1/26/2015 16:38
1/26/2015 4192 2239 9664 DESC 2/17/2015 14 16 19 24 11 10 7 1/26/2015 16:38
1/26/2015 4192 2239 9664 DESC 2/18/2015 13 15 18 21 10 9 6 1/26/2015 16:38
1/26/2015 4192 2239 9664 DESC 2/19/2015 13 15 18 24 10 8 6 1/26/2015 16:38
1/26/2015 4192 2239 9664 DESC 2/20/2015 15 18 21 26 13 11 9 1/26/2015 16:38
1/26/2015 4192 2239 9664 DESC 2/21/2015 13 16 19 24 11 9 7 1/26/2015 16:38
sqlSave(loadprod, FORECAST, tablename = "R_FORECAST", append = T,
rownames = F, colnames = FALSE, verbose = T,
safer = TRUE, addPK = FALSE,
varTypes <- tmp$TYPE_NAME,
fast = TRUE, test = FALSE, nastring = NULL)
答案 0 :(得分:1)
NUMERIC(10,10)只能保存小于1的值.NUMERIC类型定义中的第一个值是精度,即存储的总位数,第二个值是刻度,这是如何许多这些数字都在小数点之后。
TESTDB.ADMIN(ADMIN)=> create table numeric_test (col1 numeric(10,10), col2 numeric(10,0));
CREATE TABLE
TESTDB.ADMIN(ADMIN)=> insert into numeric_test values (0.1234567890, 1234567890);
INSERT 0 1
TESTDB.ADMIN(ADMIN)=> select * from numeric_test;
COL1 | COL2
--------------+------------
0.1234567890 | 1234567890
(1 row)
如果您尝试将需要小数点左侧数字的任何内容插入到COL1中,您将收到正在经历的错误。
TESTDB.ADMIN(ADMIN)=> insert into numeric_test values (10, 11);
ERROR: Numeric overflow. Value does not fit in a numeric(10,10)
如果您要查找的是存储整数值,通常应该使用INTEGER而不是NUMERIC。这提供了许多好处,包括通常更好的存储效率和对区域地图的支持。