Format the CSV file for a normal import

To import external data into the history blocks, you must format your data according to the CSV file format as outlined in the following table. For a general illustration of the format, see General file format for a CSV import.

You can name the CSV file anything you want. For the format, note that:

• Only one operation type per line is allowed.

• Multiple records per line of the same operation type is allowed.

• A multipoint update is a sequence of updates where the beginning of an update period is the end of the previous update. A multipoint update is faster than a simple sequence of inserts because a single version is used for all values. Use a multipoint update to mask underlying data with a new set of values for the specified time period.

  • Fields 3 and 4 of the values are used in single point update only and must be excluded from the record for a multipoint update. A single point update refers to the situation when an update value is assigned to a time period specified by the start date/time and end date/time. A multipoint update can replace a single or multiple previous points. It represents, like a single point update, a span of time that starts with the current row date/time and ending at the next row date/time. The value specified in each record is held as the latest value until the next record. The last record is ignored in a multipoint update.

  • The last record (time wise) will indicate the end of the previous update period. The value will be ignored.

If two multipoint update CSV files for the same tag are simultaneously copied to the \DataImport directory, the update spans across the total time for the two files. A query returning latest data hides (masks) the original version of the data from the end of the first file to the start of the second file.

For example, if the update in one file ranges from 00:00:00 to 00:05:00, and the other ranges from 00:10:00 to 00:15:00, the result is an update starting at 00:00:00 and ending at 00:15:00 ("latest"); the original data from 00:05:00 to 00:10:00 is masked as "original" data. No data is lost. To view either data from a query, use the wwVersion column to specify either "original" or "latest." By default, the latest data is shown. To prevent the masking of the original data, process the CSV files one at a time.

It is recommended not to use both inserts and original inserts for the same tag in the same file or files processed together.

When configuring the scaling setting (field 6), keep in mind that the data conversion to engineering units (a setting of 0) is performed before the value is stored. The reverse of the scaling formula for the tag is used to convert the data before storage. During retrieval, the scaling formula is applied so that the original inserted values are returned. For integer type tags, if the value after the conversion is not an integer value, it is rounded off. The rounding off can change the value to be exactly the same as the previous value, and thus the rounded off value is not stored to disk if delta storage is used. If the tag is a real type tag, the rounding off does not occur, and all values are stored.

The value to insert can be a NULL. For more information, see Handle of null values in CSV files.

For a fast load CSV import, the end time of the current block in the block80.inf file is considered to be the current time, not the current system time stamp. The end time in the .inf file is updated by the storage subsystem every 20 seconds.