从字符串插入仅插入第一行

Question

我们每月都会收到一组纯文本文件。它们没有换行符或分隔符。

我正在尝试将子字符串插入到具有大单字符串列的表中的多列和多行中。但只有第一行被插入...

CREATE TABLE [dbo].[claims_stage]
(
     [stage] [nvarchar](max) NULL
) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]

从平面文件插入到 claim_stage：

bulk insert CLAIMS_STAGE from '\\filepath\filename'

将拆分字符串从 CLAIMS_STAGE 插入新表：

insert into
CCLF8
(
    BENE_MBI_ID,
    BENE_HIC_NUM,
    BENE_FIPS_STATE_CD,
    BENE_FIPS_CNTY_CD,
    BENE_ZIP_CD,
    BENE_DOB,
    BENE_SEX_CD,
    BENE_RACE_CD,
    BENE_AGE,
    BENE_MDCR_STUS_CD,
    BENE_DUAL_STUS_CD,
    BENE_DEATH_DT,
    BENE_RNG_BGN_DT,
    BENE_RNG_END_DT,
    BENE_1ST_NAME,
    BENE_MIDL_NAME,
    BENE_LAST_NAME,
    BENE_ORGNL_ENTLMT_RSN_CD,
    BENE_ENTLMT_BUYIN_IND,
    BENE_PART_A_ENRLMT_BGN_DT,
    BENE_PART_B_ENRLMT_BGN_DT,
    BENE_LINE_1_ADR,
    BENE_LINE_2_ADR,
    BENE_LINE_3_ADR,
    BENE_LINE_4_ADR,
    BENE_LINE_5_ADR,
    BENE_LINE_6_ADR,
    GEO_ZIP_PLC_NAME,
    GEO_USPS_STATE_CD,
    GEO_ZIP5_CD,
    GEO_ZIP4_CD, 
    [INSERT_DT],
    [FILE_NAME]
)
select
    substring(stage, 1, 11)     as 'BENE_MBI_ID',
    substring(stage, 12, 11)    as 'BENE_HIC_NUM',
    substring(stage, 23, 2)     as 'BENE_FIPS_STATE_CD',
    substring(stage, 25, 3)     as 'BENE_FIPS_CNTY_CD',
    substring(stage, 28, 5)     as 'BENE_ZIP_CD',
    substring(stage, 33, 10)    as 'BENE_DOB',
    substring(stage, 43, 1)     as 'BENE_SEX_CD',
    substring(stage, 44, 1)     as 'BENE_RACE_CD',
    substring(stage, 45, 3)     as 'BENE_AGE',
    substring(stage, 48, 2)     as 'BENE_MDCR_STUS_CD',
    substring(stage, 50, 2)     as 'BENE_DUAL_STUS_CD',
    substring(stage, 52, 10)    as 'BENE_DEATH_DT',
    substring(stage, 62, 10)    as 'BENE_RNG_BGN_DT',
    substring(stage, 72, 10)    as 'BENE_RNG_END_DT',
    substring(stage, 82, 30)    as 'BENE_1ST_NAME',
    substring(stage, 112, 15)   as 'BENE_MIDLNAME',
    substring(stage, 127, 40)   as 'BENE_LAST_NAME',
    substring(stage, 167, 1)    as 'BENE_ORGNL_ENTLMT_RSN_CD',
    substring(stage, 168, 1)    as 'BENE_ENTLMT_BUYIN_IND',
    substring(stage, 169, 10)   as 'BENE_PART_A_ENRLMT_BGN_DT',
    substring(stage, 179, 10)   as 'BENE_PART_B_ENRLMT_BGN_DT',
    substring(stage, 189, 45)   as 'BENE_LINE_1_ADR',
    substring(stage, 234, 45)   as 'BENE_LINE_2_ADR',
    substring(stage, 279, 40)   as 'BENE_LINE_3_ADR',
    substring(stage, 319, 40)   as 'BENE_LINE_4_ADR',
    substring(stage, 359, 40)   as 'BENE_LINE_5_ADR',
    substring(stage, 399, 40)   as 'BENE_LINE_6_ADR',
    substring(stage, 439, 100)  as 'GEO_ZIP_PLC_NAME',
    substring(stage, 539, 2)    as 'GEO_USPS_STATE_CD',
    substring(stage, 541, 5)    as 'GEO_ZIP5_CD',
    substring(stage, 546, 4)    as 'GEO_ZIP4_CD',
    GETDATE()                   as 'INSERT_DT',
    'filename'                  as 'FILE_NAME'
from 
    CLAIMS_STAGE

我应该循环，我如何循环？

Answer 1

这是解析数据的一种方法。它需要一个数字/计数表 - 这是一个永久表，其中包含从 1 开始的单个 int 列，最多可达 bazillion 或您需要的任何值。

这里我只是为示例动态创建它：

/* Sample data, 3 "rows" of fixed-width data as a single row in table */

create table claims_stage (stage varchar(max))
insert into claims_stage
select Cast('  123' as char(5)) + cast('Homer' as Char(10)) + Cast('Simpson' as char(10)) + Cast('  456' as char(5)) + 
       Cast('   56' as char(5)) + cast('Bart' as Char(10)) + Cast('Simpson' as char(10)) + Cast('  888' as char(5)) +
       Cast('  122' as char(5)) + cast('Stuey' as Char(10)) + Cast('Griffin' as char(10)) + Cast('  60' as char(5))


with d as (select * from (values (0),(1),(2))n(n)), /* Tally table */
CCLF8 as (
  select Substring(stage, 1 + d.n * 30, 30) stage           /* Knowing the length of each row, */
  from claims_stage s                                 /* break it into multiple rows */
  cross join d
)

现在从中选择我们的列

select 
    Substring(stage,1,5) Col1,
    Substring(stage,6,10) Col2,
    Substring(stage,16,10) Col3,
    Substring(stage,25,5) Col4
from CCLF8