|
|
This pop-up form imports data from various formats and will create a
ring of data in a datastore. This data is then native to habitat
and can be used in the same way as information collected from clockwork
Step 1 Type in the source file or click '...' to browse
|
 |
Step 2 Choose tab or comma separated data format or
sar the well known unix monitoring tool
|
Step 3 Select name and/or information rows
above each column (see below)
|
Step 4 Select time and sequence columns in addition to data
(see below)
|
Step 5 Type in the file name of the destination data store
or click the `...' button to browse
|
Step 6 Type in the ring name where the imported data
should be stored in the file.
|
Step 7 Check the details, then click Import to
run the import process; errors will appear above the buttons.
When you have done, click Finished
|
column names |
A row of names is expected as the first line in a comma or tab
seperated format. These headers will form the column names for the
subsequent data. If this option is not slected, then unique column
names will be generated for you.
tom dick harry
1 2 3
4 5 6
|
column information |
If ticked, information rows are expected in the import file in
the same format that habitat exports them.
These rows contain additional, non-data information about each
column, such as an enhanced description or a data type identifier.
The importer expects zero or more rows between the header
(if selected) and a single line containing two dashes, `--'.
As well as column Information, an additional
unlabeled column is printed at the end of each line to name the
information row.
This is unusual and is only likely to come from a habitat
exported file.
tom dick harry
thomas richard henry alternate
--
1 2 3
4 5 6
In the above example, we see an info line called alternate
giving more information about each column.
|
insertion time |
An additional column will be expected in the import file,
named inserted, that contains the time of the samples.
When imported, the sample will have its time associcated with it
instead of the insertion time, which is the normal behaviour.
The type of data is time_t in C-speak, which is the number of
seconds since the unix epoch, Janurary 1 1970. Clearly, if column
names are absent and the option is not ticked, the importer will
not be able to recognise a time column.
inserted tom dick harry
998337060 1 2 3
998337120 4 5 6
|
sequence number |
An additional column will be expected in the import file,
named sequence, that contains a positive integer that orders
and groups samples together.
If multiple, seperate samples have the same time value (they arrived
at the sampling agent during the same second), the sequence number
preserves the order of data.
Secondly, the sequence groups multi line samples together, useful in
processing the same attributes from multiple instances like disk
statistics.
The number will not recorded, but the order and grouping
will be preserved. If the the option is not ticked but sample time
is available, then the time will be used as an an effective
sequence ordering, with samples sharing the same time assumed to
be a multi-line sample. If neither sequence or time is available
all rows are assumed to be single line samples in sequence order.
inserted sequence read/s write/s disk
998337060 0 1 2 1
998337060 0 4 5 2
998337060 1 7 8 1
998337060 1 10 11 2
The table above shows two samples over four rows. As the
inserted time column is the same, only the two values from
sequence tells us there are two sets of samples.
In this case disk changes and will be the instance.
Two samples of two rows each will be created.
|
|