amazon-redshift - 列是没有时区的时间戳类型，但表达式的类型是字符变化

Question

我正在尝试将数据从一个表（在 Redshift 中）移动到另一个表（在 Redshift 中）。sql 代码使用 INSERT INTO SELECT 并且不知何故它不起作用。两个表中列的数据类型不同，但在我的 SELECT 子句中，我使用 SQL UDF 将其转换为各自的数据类型。

出于调试目的，我尝试使用 UPLOAD 命令在 S3 中上传仅包含一列（Redhsift 错误输出的列）的表，并使用 COPY 命令将其复制到 Redshift 中的表中。现在，当我尝试使用这个只有 1 列的表，使用 UDF 插入 INTO SELECT 时，它工作正常。

以下是表定义目标表：

CREATE TABLE test.dn_ems_activity_kush
(
    tstamp_trans TIMESTAMP WITHOUT TIME ZONE ENCODE ZSTD
    ,tstamp_local VARCHAR(25) ENCODE RAW
    ,pan VARCHAR(131) ENCODE ZSTD
    ,retrieval_ref_no VARCHAR(12) ENCODE ZSTD
    ,case_type_ind VARCHAR(2) ENCODE ZSTD
    ,sys_trace_audit_no VARCHAR(6) ENCODE ZSTD
    ,case_no VARCHAR(14) ENCODE ZSTD
    ,net_id_ems VARCHAR(3) ENCODE ZSTD
    ,net_term_id VARCHAR(8) ENCODE ZSTD
    ,rpt_lvl_id_b VARCHAR(10) ENCODE ZSTD
    ,status VARCHAR(4) ENCODE ZSTD
    ,request_type VARCHAR(4) ENCODE ZSTD
    ,state_tstamp TIMESTAMP WITHOUT TIME ZONE ENCODE ZSTD
    ,state_expir_tstamp TIMESTAMP WITHOUT TIME ZONE ENCODE ZSTD
    ,acct_id_1 VARCHAR(28) ENCODE ZSTD
    ,amt_recon_net NUMERIC(16,2) ENCODE ZSTD
    ,amt_tran NUMERIC(16,2) ENCODE ZSTD
    ,cur_recon_net VARCHAR(3) ENCODE ZSTD
    ,cur_tran VARCHAR(3) ENCODE ZSTD
    ,card_acpt_city VARCHAR(27) ENCODE ZSTD
    ,card_acpt_name VARCHAR(25) ENCODE ZSTD
    ,card_acpt_term_id VARCHAR(15) ENCODE ZSTD
    ,inst_id_acq VARCHAR(11) ENCODE ZSTD
    ,inst_id_iss VARCHAR(11) ENCODE ZSTD
    ,merchant_cat_code VARCHAR(4) ENCODE BYTEDICT
    ,priority VARCHAR(1) ENCODE ZSTD
    ,proc_id_acq VARCHAR(8) ENCODE ZSTD
    ,proc_id_iss VARCHAR(8) ENCODE ZSTD
    ,tran_type_id VARCHAR(10) ENCODE ZSTD
    ,case_group VARCHAR(4) ENCODE ZSTD
    ,cashback_amt NUMERIC(16,2) ENCODE ZSTD
    ,cur_cashback VARCHAR(3) ENCODE ZSTD
    ,account_type VARCHAR(4) ENCODE ZSTD
    ,amt_adjustment NUMERIC(16,2) ENCODE ZSTD
    ,cur_adjustment VARCHAR(3) ENCODE ZSTD
    ,action_to_cardhldr VARCHAR(1) ENCODE ZSTD
    ,adjustment_reason VARCHAR(254) ENCODE ZSTD
    ,reason_code VARCHAR(4) ENCODE BYTEDICT
    ,fraud_reason VARCHAR(1) ENCODE ZSTD
    ,tstamp_created TIMESTAMP WITHOUT TIME ZONE ENCODE ZSTD
    ,transition_user_id VARCHAR(8) ENCODE ZSTD
    ,request_type_prev VARCHAR(4) ENCODE ZSTD
    ,status_prev VARCHAR(4) ENCODE ZSTD
    ,request_type_next VARCHAR(4) ENCODE ZSTD
    ,status_next VARCHAR(4) ENCODE ZSTD
    ,accounting_type VARCHAR(5) ENCODE ZSTD
    ,accounting_user_id VARCHAR(8) ENCODE ZSTD
    ,debit_acct_id VARCHAR(28) ENCODE ZSTD
    ,credit_acct_id VARCHAR(28) ENCODE ZSTD
    ,amount_val NUMERIC(16,2) ENCODE ZSTD
    ,switch_name VARCHAR(10) ENCODE ZSTD
    ,assgn_id VARCHAR(20) ENCODE ZSTD
    ,src_file_name VARCHAR(250) ENCODE ZSTD
    ,load_time TIMESTAMP WITHOUT TIME ZONE ENCODE ZSTD
)
DISTSTYLE KEY
 DISTKEY (pan)
 SORTKEY (
    tstamp_local
    );

源表：

CREATE TABLE test.padb_dn_ems_activity_history_kush
(
 tstamp_trans character varying(16)
,tstamp_local character varying(14)
,pan character varying(131)
,retrieval_ref_no character varying(12)
,case_type_ind character varying(2)
,sys_trace_audit_no character varying(6)
,case_no character varying(14)
,net_id_ems character varying(3)
,net_term_id character varying(8)
,rpt_lvl_id_b character varying(10)
,status character varying(4)
,request_type character varying(4)
,state_tstamp character varying(16)
,state_expir_tstamp character varying(16)
,acct_id_1 character varying(28)
,amt_recon_net numeric(18,2)
,amt_tran numeric(18,2)
,cur_recon_net character varying(3)
,cur_tran character varying(3)
,card_acpt_city character varying(27)
,card_acpt_name character varying(25)
,card_acpt_term_id character varying(15)
,inst_id_acq character varying(11)
,inst_id_iss character varying(11)
,merchant_cat_code character varying(4)
,priority character varying(1)
,proc_id_acq character varying(8)
,proc_id_iss character varying(8)
,tran_type_id character varying(10)
,case_group character varying(4)
,cashback_amt numeric(20,2)
,cur_cashback character varying(3)
,account_type character varying(4)
,amt_adjustment numeric(18,2)
,cur_adjustment character varying(3)
,action_to_cardhldr character varying(1)
,adjustment_reason character varying(254)
,reason_code character varying(4)
,fraud_reason character varying(1)
,tstamp_created character varying(16)
,transition_user_id character varying(8)
,request_type_prev character varying(4)
,status_prev character varying(4)
,request_type_next character varying(4)
,status_next character varying(4)
,accounting_type character varying(5)
,accounting_user_id character varying(8)
,debit_acct_id character varying(28)
,credit_acct_id character varying(28)
,amount_val numeric(18,2)
,switch_name character varying(10)
);

函数定义：

create or replace
function fdatetime(
    varchar(20)
) returns timestamp stable as $$ select
    case
        when length($1) > 14 then cast(to_timestamp(substring($1, 1, 4)+ '-' + substring($1, 5, 2)+ '-' + substring($1, 7, 2)+ ' ' + substring($1, 9, 2)+ ':' + substring($1, 11, 2)+ ':' + substring($1, 13, 2)+ '.' + substring($1, 15, length($1)-14), 'YYYY-MM-DD HH24:MI:SS:MS') as timestamp)
        when length($1) = 14 then cast(to_timestamp(substring($1, 1, 4)+ '-' + substring($1, 5, 2)+ '-' + substring($1, 7, 2)+ ' ' + substring($1, 9, 2)+ ':' + substring($1, 11, 2)+ ':' + substring($1, 13, 2), 'YYYY-MM-DD HH24:MI:SS:MS') as timestamp)
        when length($1) = 8 then cast(to_timestamp(substring($1, 1, 4)+ '-' + substring($1, 5, 2)+ '-' + substring($1, 7, 2), 'YYYY-MM-DD HH24:MI:SS:MS') as timestamp)
        else null
    end $$ language sql;

最后的 SELECT INTO INSERT 语句：

INSERT INTO
  test.dn_ems_activity_kush(
    tstamp_trans,
    tstamp_local,
    pan,
    retrieval_ref_no,
    case_type_ind,
    sys_trace_audit_no,
    case_no,
    net_id_ems,
    net_term_id,
    rpt_lvl_id_b,
    status,
    request_type,
    state_tstamp,
    state_expir_tstamp,
    acct_id_1,
    amt_recon_net,
    amt_tran,
    cur_recon_net,
    cur_tran,
    card_acpt_city,
    card_acpt_name,
    card_acpt_term_id,
    inst_id_acq,
    inst_id_iss,
    merchant_cat_code,
    priority,
    proc_id_acq,
    proc_id_iss,
    tran_type_id,
    case_group,
    cashback_amt,
    cur_cashback,
    account_type,
    amt_adjustment,
    cur_adjustment,
    action_to_cardhldr,
    adjustment_reason,
    reason_code,
    fraud_reason,
    tstamp_created,
    transition_user_id,
    request_type_prev,
    status_prev,
    request_type_next,
    status_next,
    accounting_type,
    accounting_user_id,
    debit_acct_id,
    credit_acct_id,
    amount_val,
    switch_name
  )
SELECT
  fdatetime(tstamp_trans),
  fdatetimeintostring(tstamp_local),
  pan,
  retrieval_ref_no,
  case_type_ind,
  sys_trace_audit_no,
  case_no,
  net_id_ems,
  net_term_id,
  rpt_lvl_id_b,
  status,
  request_type,
  fdatetime(state_tstamp),
  fdatetime(state_expir_tstamp),
  acct_id_1,
  ROUND(amt_recon_net / 100, 2),
  Round(amt_tran / 100, 2),
  cur_recon_net,
  cur_tran,
  card_acpt_city,
  card_acpt_name,
  card_acpt_term_id,
  inst_id_acq,
  inst_id_iss,
  merchant_cat_code,
  priority,
  proc_id_acq,
  proc_id_iss,
  tran_type_id,
  case_group,
  Round(cashback_amt / 100, 2) cashback_amt,
  cur_cashback
  account_type,
  Round(amt_adjustment / 100, 2),
  cur_adjustment,
  action_to_cardhldr,
  adjustment_reason,
  reason_code,
  fraud_reason,
  fdatetime(tstamp_created),
  transition_user_id,
  request_type_prev,
  status_prev,
  request_type_next,
  status_next,
  accounting_type,
  accounting_user_id,
  debit_acct_id,
  credit_acct_id,
  Round(amount_val / 100, 2),
  switch_name
FROM
  test.padb_dn_ems_activity_history_kush;

SQL 错误 [500310] [42804]：Amazon 无效操作：列“tstamp_created”的类型是没有时区的时间戳，但表达式的类型是字符变化；

标签： amazon-redshift

Answer 1

这只是一个语法错误。在 column 之后的 select 子句中cur_cashback，我忘记放置逗号。这就是出现错误的原因。