exceptionfactory commented on code in PR #9874:
URL: https://github.com/apache/nifi/pull/9874#discussion_r2075680791
##########
nifi-extension-bundles/nifi-poi-bundle/nifi-poi-services/src/main/java/org/apache/nifi/excel/RowIterator.java:
##########
@@ -42,13 +51,26 @@ class RowIterator implements Iterator<Row>, Closeable {
private Row currentRow;
RowIterator(final InputStream in, final ExcelRecordReaderConfiguration
configuration, final ComponentLog logger) {
- this.workbook = StreamingReader.builder()
- .rowCacheSize(100)
- .bufferSize(4096)
- .password(configuration.getPassword())
- .setAvoidTempFiles(configuration.isAvoidTempFiles())
- .setReadSharedFormulas(true) // NOTE: If not set to true, then
data with shared formulas fail.
- .open(in);
+ if (isXSSFExcelFile(in, configuration.getPassword())) {
+ this.workbook = StreamingReader.builder()
+ .rowCacheSize(100)
+ .bufferSize(4096)
+ .password(configuration.getPassword())
+ .setAvoidTempFiles(configuration.isAvoidTempFiles())
+ .setReadSharedFormulas(true) // NOTE: If not set to true,
then data with shared formulas fail.
+ .open(in);
+ } else {
+ // Providing the password to the HSSFWorkbook is done by setting a
thread variable managed by
+ // Biff8EncryptionKey. After the workbook is created, the thread
variable can be cleared.
+
Biff8EncryptionKey.setCurrentUserPassword(configuration.getPassword());
Review Comment:
@zhtk The thread safety issue is that calling a static method applies to all
instances of the Reader. That means if there are multiple instances of the
Excel Reader configured, with different passwords, different threads could step
on each other. I recommend taking a closer look at the POI library to see if
there are any other options. If this is the only approach supported in the
library itself, then it would require more significant custom code to implement
a solution that works in the context of NiFi.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
