Database Reference
In-Depth Information
a line number. This is more convenient when dealing with large files because
it does not require reading the file from the beginning to count lines.
The most interesting error is the one that appears in the
errorResult
field. It does not correspond to any of the errors that were encountered while
parsing the data.
{
"reason": "invalid",
"message": "Too many errors encountered. Limit is: 0."
}
The job did not fail due to the individual parse errors that resulted in
bad records. Rather the job failed because the number of bad records
encountered exceeded a threshold. The default limit for the number of bad
records allowed in a load job is zero. This limit can be modified, but first
you should discuss why you might want to change its value. Generally, when
loading data into the service, users expect that all the data will be loaded
without any errors. If errors are present it usually indicates a serious error in
how the data is being supplied, which needs to be corrected before retrying
the load operation. However, in some cases users are dealing with imperfect
data; for example, the collection process corrupts a small number of records.
If the data is still useful with these corrupt records dropped, it is reasonable
to have the service ignore some number of faulty records. To support this
load jobs accept a
maxBadRecords
parameter. For example, modifying the
load job you just ran to set this parameter can make the job succeed:
loadConfig['maxBadRecords'] = 7
The job still reports the bad records encountered, but if they are fewer than
the value specified in this parameter, the job still succeeds and all the good
records are added to the destination table. If there are a large number of
bad records, only a sample of them will be reported in the
errors
list;
however, all the bad records are counted for the purpose of enforcing the
maxBadRecords
test.
This completes the review of the errors commonly encountered by load jobs.
One class of errors mentioned but not discussed in detail is quota errors.
The next section discusses the limits that apply to load jobs and also covers
the related errors raised when these limits are exceeded.
