Errata

indentation in the function definition seems to be wrong.

Note from the Author or Editor:
Correct. It should look like:
```
def map_rows(self, json_payload: str) -> str:
try:
evt = json.loads(json_payload)
evt_time = int(datetime.datetime.fromisoformat(evt['event_time']).
yield json.dumps({'visit_id': evt['visit_id'], 'event_time': evt_time,
'page': evt['page']})
except Exception as e:
yield self.invalid_data_output, _wrap_input_with_error(json_payload, e)
```

Thank you!

The list does not match the figure 5-4.
the first 2 items seem to be flipped.

Note from the Author or Editor:
Thank you. Yes, it should be changed to:

"""
Implementation 1 stores computed rows as a single flat structure and the original row as nested attributes.
• Implementation 2 does the opposite (i.e., it stores the original row in a flat structure and all computed columns as nested attributes.).
"""

"The first two implementations use a denormalization approach that may be faster at reading. The third one uses the normalized approach, "

Actually there are four implementation and the first "three" implementations
are denormalized approach, and the "fourth" one is normalized approach.

Note from the Author or Editor:
Exactly, it should be:
```
The first three implementations use a denormalization approach that may be faster at reading. The fourth one uses the normalized approach
```

indentation of the last two lines is strange. I guess one more space is required.

Note from the Author or Editor:
Exactly, we should add an extra whitespace to the last 2 lines so that it becomes aligned with the ifs:

```
if len(delivery_groups) <= groups_index:
delivery_groups.append([])
delivery_groups[groups_index].append(visit)
groups_index += 1
```

indentation of the last line is wrong.

Note from the Author or Editor:
Yes, we should remove the indentation so that it becomes:
```
clear_context >> file_sensor >> visits_loader >> generate_cube
```
(no whitespaces in the beginning)

'Unfortunately, this technique won’t work for the Parallel Split.'

'Parallel split' should be 'Exclusive Choice'.

Note from the Author or Editor:
Exactly, it should be "Exclusive Choice"; Thank you!

Indentation is confusing.

Note from the Author or Editor:
We might add some indentation to this part:
```
countries_area_mapping = {
'Poland': 'eu', 'France': 'eu', 'Spain': 'eu', 'the USA': 'na'
}
```

But overall it looks quite ok.

name_w_version should be name_with_version.

I guess this is due to page width limitation,
but it have to be consistent with two other occurrence in the
same code.

Note from the Author or Editor:
Exactly, we should change the "name_with_version" in this example to "name_w_version"


> I guess this is due to page width limitation,
Thank you! And exactly, I changed it afterwards to handle page limitation and didn't double check the code.

start_date_to_backfill = (dr.execution_date
datetime.timedelta(days=days_to_backfill))

`-` is required between these lines.

Note from the Author or Editor:
Yes, exactly. It should be:
```
start_date_to_backfill = (dr.execution_date -
datetime.timedelta(days=days_to_backfill))
```

version_to_restore = current_run_version - 1
(DeltaTable.forName(spark, 'devices').restoreToVersion(previous_job_version ))

'previous_job_version' should be 'version_to_restore'

Note from the Author or Editor:
Correct, thank you!

It should be then:
```
(DeltaTable.forName(spark, 'devices').restoreToVersion(version_to_restore))
```

new_version_df = (spark.createDataFrame([
Row(execution_time=current_execution_time, delta_table_version=last_version)]))

'new_version_df' should be 'new_version', since 'new_version_df' is not referred afterward.

Note from the Author or Editor:
Thank you, yes!

new_version_df = (spark.createDataFrame([

should become

new_version = (spark.createDataFrame([

"Apache Parquet statistics for the login column"

As the text above says, it shows ID column, instead of login.

Note from the Author or Editor:
Thank you. Yes, it should be:

Apache Parquet statistics for the ID column