admin管理员组文章数量:1025217
I am facing a significant challenge.
I have several procedures in my database that export data from queries using UTL_FILE. However, UTL_FILE has a limitation of 32KB per file, which prompted me to come up with a solution: I implemented a loop where, in each iteration, the .txt file is incremented with data no larger than 32KB. This approach worked well for a while, but after modifying my queries, the data size increased to over 10 million lines. As a result, my original looping strategy has become impractical, as it requires running the SQL query, creating the .txt file, saving up to 32KB of data, and closing the file during each iteration.
Considering that my query only takes about 5 minutes to execute, the time required for each loop becomes excessive, especially when I have to run 50 to 100 iterations. This means each iteration takes at least 5 minutes, multiplied by the number of loops, leading to hours of processing time.
Currently, I have 9 queries, and the total export process takes approximately 8 to 9 hours. I need to find a way to significantly reduce this time. For context, I created an export program in Python, and the same data exports take less than 1 hour.
Could you suggest a faster alternative to export large amounts of data into .txt files from Oracle?
Example:
PROCEDURE EMPLOYEE IS
limite INTEGER;
counter INTEGER := 0;
step INTEGER := 500;
file UTL_FILE.FILE_TYPE;
name_file VARCHAR(50) := 'X_TEST_' || TO_CHAR(sysdate, 'YYYYMMDDHH24MISS')||'_fornecedores.txt';
BEGIN
file := UTL_FILE.FOPEN(location => 'MYDIR',
filename => name_file,
open_mode => 'w');
SELECT COUNT(*)
INTO limite
FROM table f
LEFT JOIN table_2 c
ON f.seqfornecedor = c.seqfornecedor
AND c.indprincipal = 'S'
JOIN table_3 g
ON f.seqfornecedor = g.seqpessoa
WHERE f.statusgeral = 'A'
AND f.dtaalteracao = TRUNC(sysdate - 1);
WHILE counter <= limite LOOP
file := UTL_FILE.FOPEN(location => 'MYDIR',
filename => name_file,
open_mode => 'a');
FOR k IN (
SELECT desc_fornec,
cod,
fantasia,
cpf_cnpj,
contato,
emailcontato,
fixo_1,
fixo_2
FROM (SELECT g.seqpessoa || '_' || g.nomerazao AS desc_fornec,
g.seqpessoa AS cod,
g.fantasia,
CASE
WHEN g.fisicajuridica = 'J' THEN
LPAD(g.nrocgccpf, 12, '0') || LPAD(g.digcgccpf, 2, '0')
WHEN g.fisicajuridica = 'F' THEN
LPAD(g.nrocgccpf, 9, '0') || LPAD(g.digcgccpf, 2, '0')
END as CPF_CNPJ,
f.nomecontato AS contato,
f.emailcontato,
1 AS FIXO_1,
0 AS FIXO_2,
ROW_NUMBER() OVER(ORDER BY g.seqpessoa) AS row_num
FROM table f
LEFT JOIN table_2 c
ON f.seqfornecedor = c.seqfornecedor
AND c.indprincipal = 'S'
JOIN table_3 g
ON f.seqfornecedor = g.seqpessoa
WHERE f.statusgeral = 'A'
AND f.dtaalteracao = TRUNC(sysdate - 1))
WHERE row_num > counter
AND row_num <= counter + step
)
LOOP
UTL_FILE.PUTF(file,
k.desc_fornec || ';' ||
k.cod || ';' ||
k.fantasia || ';' ||
K.CPF_CNPJ || ';' ||
k.contato || ';' ||
k.emailcontato || ';' ||
k.fixo_1 || ';' ||
k.fixo_2 || UTL_TCP.CRLF);
END LOOP;
UTL_FILE.FCLOSE(file);
counter := counter + step;
END LOOP;
END;
I am facing a significant challenge.
I have several procedures in my database that export data from queries using UTL_FILE. However, UTL_FILE has a limitation of 32KB per file, which prompted me to come up with a solution: I implemented a loop where, in each iteration, the .txt file is incremented with data no larger than 32KB. This approach worked well for a while, but after modifying my queries, the data size increased to over 10 million lines. As a result, my original looping strategy has become impractical, as it requires running the SQL query, creating the .txt file, saving up to 32KB of data, and closing the file during each iteration.
Considering that my query only takes about 5 minutes to execute, the time required for each loop becomes excessive, especially when I have to run 50 to 100 iterations. This means each iteration takes at least 5 minutes, multiplied by the number of loops, leading to hours of processing time.
Currently, I have 9 queries, and the total export process takes approximately 8 to 9 hours. I need to find a way to significantly reduce this time. For context, I created an export program in Python, and the same data exports take less than 1 hour.
Could you suggest a faster alternative to export large amounts of data into .txt files from Oracle?
Example:
PROCEDURE EMPLOYEE IS
limite INTEGER;
counter INTEGER := 0;
step INTEGER := 500;
file UTL_FILE.FILE_TYPE;
name_file VARCHAR(50) := 'X_TEST_' || TO_CHAR(sysdate, 'YYYYMMDDHH24MISS')||'_fornecedores.txt';
BEGIN
file := UTL_FILE.FOPEN(location => 'MYDIR',
filename => name_file,
open_mode => 'w');
SELECT COUNT(*)
INTO limite
FROM table f
LEFT JOIN table_2 c
ON f.seqfornecedor = c.seqfornecedor
AND c.indprincipal = 'S'
JOIN table_3 g
ON f.seqfornecedor = g.seqpessoa
WHERE f.statusgeral = 'A'
AND f.dtaalteracao = TRUNC(sysdate - 1);
WHILE counter <= limite LOOP
file := UTL_FILE.FOPEN(location => 'MYDIR',
filename => name_file,
open_mode => 'a');
FOR k IN (
SELECT desc_fornec,
cod,
fantasia,
cpf_cnpj,
contato,
emailcontato,
fixo_1,
fixo_2
FROM (SELECT g.seqpessoa || '_' || g.nomerazao AS desc_fornec,
g.seqpessoa AS cod,
g.fantasia,
CASE
WHEN g.fisicajuridica = 'J' THEN
LPAD(g.nrocgccpf, 12, '0') || LPAD(g.digcgccpf, 2, '0')
WHEN g.fisicajuridica = 'F' THEN
LPAD(g.nrocgccpf, 9, '0') || LPAD(g.digcgccpf, 2, '0')
END as CPF_CNPJ,
f.nomecontato AS contato,
f.emailcontato,
1 AS FIXO_1,
0 AS FIXO_2,
ROW_NUMBER() OVER(ORDER BY g.seqpessoa) AS row_num
FROM table f
LEFT JOIN table_2 c
ON f.seqfornecedor = c.seqfornecedor
AND c.indprincipal = 'S'
JOIN table_3 g
ON f.seqfornecedor = g.seqpessoa
WHERE f.statusgeral = 'A'
AND f.dtaalteracao = TRUNC(sysdate - 1))
WHERE row_num > counter
AND row_num <= counter + step
)
LOOP
UTL_FILE.PUTF(file,
k.desc_fornec || ';' ||
k.cod || ';' ||
k.fantasia || ';' ||
K.CPF_CNPJ || ';' ||
k.contato || ';' ||
k.emailcontato || ';' ||
k.fixo_1 || ';' ||
k.fixo_2 || UTL_TCP.CRLF);
END LOOP;
UTL_FILE.FCLOSE(file);
counter := counter + step;
END LOOP;
END;
本文标签： plsqlExport txt from Oracle DBStack Overflow
版权声明：本文标题：plsql - Export .txt from Oracle DB - Stack Overflow 内容由热心网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://it.en369.cn/questions/1745618136a2159413.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。
发表评论