Importing daily data

[rdstern]

@mhabimana three more questions about importing daily data:

Your file in the example in Section 8.5 had columns that all correspond with those you could match with the field names in the column headers. a) Do we need a column with the hours? b) Can we have extra columns? We also had a column, called station, with the station names? Does that cause a problem If we can, then what do we do with it? Just leave it out, when we attach the column headers? c) We quite often had data with duplicate days, i.e. same station, same element, same date. What happens then?
d) How do we put trace rainfall in?

[mhabimana]

@rdstern Thanks a) This will depend with your data,if your data has hours,the response is yes,if not,you don't need this(and therefore no need to match this). In short,you need to match according to your data structure columns;

b) If you have a station name column,don't match it,leave it the way it is and match other columns ,the stationid is the most important (i.e. column which are not matched with corresponding elements will not be imported. since you have already defined the metadata for that station,you don't need a column for station,the stationid is enough;

c) Data importation will start after clicking on the button Load Data,these data will go straightforward into observationinitial and since this table does not accept duplicates,if you have transferred two similar records,one will be picked and one left aside;

d) since this is external data,trace should be 0 and not 0T as in data entry forms(so,not flag in this case).
Sorry for my previous answer sent to your mail box.

[rdstern]

Many thanks. My only further query is on b) i.e. extra columns. We made a mistake once and made it ID and then realised that this was not the ID. We could not find an easy way to take away the information. I wonder if there should also be an extra field in the list - which is blank - so you can easily cater for extra column s even if you first matched them, and then needed to correct.

[mhabimana]

@rdstern Thanks,yes it is possible .I am sure @smachua is following the discussion.However,under the administration,update observations,you can be able to select observationinitial,then, view the station with errors and delete rows with issues or update.It is possible to select and delete many rows at once. The only option to correct is via the update observations facility and here you can delete/edit and then import a fresh.

[rdstern]

Great thanks - that is the right place to use, i.e. Administration > Update Observations - and there is a view button there, which is fine.

Now all I would like there is for the Station field to be a pull-down with the station ID and name. In each table it just gives the stations that have data in that table.

Similarly for the element information. That becomes a pull-down. If you select a station, then it shows the elements in the table for that station. If the Station field is blank, then it shows all the elements anywhere in that table.

To try - with the test database final table, I left everything blank and hoped to have a view of all the data - or at least a message saying 72000 records - too many to view, please make a selection. It said I had no data instead.

Currently it is a bit painful to use, because I have to remember station IDs. It is also very limited, because (I think) it only works when all 4 fields, Station, Element, Start year, End Year are filled.

Surely if I leave the start and end years blank it could give all the data. And when Element is blank it could give the data for all elements.

And when it can't find any data it could do much better than the current all-embracing message. If the station has no data for that element, then it could (for example)) say what elements are at that station.

This could be a great dialogue - it is one I have been looking for.

I would argue that enhancing this dialogue is urgent. Not for Rwanda, but for the next update.

[mhabimana]

@rdstern Well noted.

[mhabimana]

@smachua : Under this issue,@rdstern suggested that the selected text boxes below be converted to drop down lists to avoid users typing errors.

[smachua]

This dialog is meant for observations update (mainly corrections and deletions) . By choosing the dialog the administrator knows exactly which data to work ok. To ensure that no data is selected by mistake and erroneously changed list selections was avoided by design. I did explain all these facts to Roger in Rwanda.