Error on merge

Dear All,





I am trying to import a tab delimited file (from excel) into an existing database. This is only text data (no structures) I am using InstantJchem version 2 downloaded this morning (M3) on linux.





The data will not import, giving the following errors.





code_nospace is mapped to current field code_nospace


active is mapped to current field active


active2 is mapped to current field active2


Confirmed is mapped to current field Confirmed


Starting to import data...


Error: Failed to import row 1


Empty structures not allowed


Error: Failed to import row 2


Empty structures not allowed


Error: Failed to import row 3


Empty structures not allowed


Error: Failed to import row 4


Empty structures not allowed


Error: Failed to import row 5


Empty structures not allowed


Error: Failed to import row 6


Empty structures not allowed


Error: Failed to import row 7


Empty structures not allowed


Error: Failed to import row 8


Empty structures not allowed


Error: Failed to import row 9


Empty structures not allowed


Error: Failed to import row 10


Empty structures not allowed


Error: Failed to import row 11


Empty structures not allowed


Error:





TOO MANY CONSECUTIVE ERRORS, TERMINATING IMPORT





Import completed in 0s.


10 rows succesfully imported.


11 Errors.





Do you have a suggestion as to what is wrong?





Cheers





David

A reply to my own e-mail.





I have updated InstantJchem (v 2) using the update centre.





I found the "Empty structure allowed" box in the database schema which now fixes the prevous error (This is not very obvious and the database entries do, in fact, already have structures, they are just not being imported from the excel file).





The data import then seems to run OK (see below) except that no data is actually imported. ie new columns are created in the database but they are all empty.





Regards





David





Output:


----------





Created new field active


Created new field active2


Created new field Confirmed


code_nospace is mapped to current field code_nospace


Starting to import data...


Import completed in 1s.


500 rows succesfully imported.


0 Errors.

Hi David,





With your first problem are you sure you are specifying the merge fields? The log that you included only mentioned mapped fields, not a merge field. If no merge fields are specified the data will be imported as new rows, and so will (correctly) fail if your table does not allow empty structures.


To specify a field for the merge you would typically need to remove it from the currently mapped fields, then select it in both list of fields and click on the merge button. The symbol next to the field in the right hand panel will change to the double headed arow signifying that the field is being used for merging.





With your second problem with the data not apparantly being updated, I think this is because of a know bug that results in the display not being updated after the import has compelted. The data has actually been properly loaded, but you are just not seeing it. Closing and restarting IJC will make the data appear.


We are working on fixing this bug.


Can you try this and let me know if it turns out to be somehing different.








Tim

Hello all,





I have installed the last IJC 2.0 milestone release 3.a. When I import a txt file into an existing structure table, I noticed the following issues:


- all the txt headers of my text file are renamed field1, field2, … and are imported with empty cells. I supposed that maybe this was due to the decimal sign (comma) instead of dots, but modifying this didn’t solve the problem.


- In import step 3 “Field details”, imported fields are inactive as well as all possible operations (Add, Merge…)


- When the import is done using a csv file, only one field is imported.





Isi

Quote:

all the txt headers of my text file are renamed field1, field2

By default the first line is treated as data. If you want it to be treated as field names then you need to set the parser options (click on the ... button next to the 'File type' selector and check the 'First line is field names' option.

Quote:

and are imported with empty cells

This is a know bug. The display does not update but the data is there. If you shutdown and restart then you should see it.


We are working on fixing this bug.

Quote:

In import step 3 “Field details”, imported fields are inactive as well as all possible operations (Add, Merge…)

Fields that are already specified for import will be inactive - you can't import the same field twice. To re-specify how the field is handled remove it from the right hand list, and then select it from the left hand list (it will now be active) and specify how you want to handle it (e.g map, merge).

Quote:

When the import is done using a csv file, only one field is imported.

The file type was not detected as CSV (we will fix this). In the parser options you can force this to be treated as CSV and all should be OK.





Tim

Many thanks, it works!





Best regards,





Isi

Hi Tim,





Thanks for you help. It all works now.





Perhaps the main problem is that I didn't find the merge menu particularly intuitive. ie. The single and double headed arrows implying add and merge do not mean anything to me. It is also not clear that you should first remove a mapped field then add it again.





Another minor comment. The 'file import' wizard does not recognise .csv file format by file extension, it assumes that all delimited files are tab delimited. It would be nice if it did this.





Cheers





David

Hi David,





One big problem in the import UI is getting all the possible options onto the screen in a reasonably compact, but simple manner, and without making the basic "import SD file" (which is what will be used 95% of the time) any more complex than it has to be.


If you have suggestions for improving this please let us know e.g if you have idea for better icons etc.





We know about the failure to recognise the cvs format files. We will fix this for the next release.





Thanks





Tim

Hi Tim,





I am not really a fan of icons for complex concepts they do OK for 'print', 'home' and 'reload' (to look at my web browser) but I think that they often fail to be useful when used for more complex concepts.





What about a drop down 'button' for each field in the database to replace the + and arrow symbols? It could have the options 'ignore', 'add' and 'merge'. A second drop down button could be used to map fields from the input file to fields in the databse?





Does that make sense?





Cheers





David