Re: "Ungroup" data for import into PostgreSQL

Sorry for the late reply...life interefered...

From: "Jim Nasby" <Jim.Nasby@BlueTreble.com>


> On 1/15/15 9:43 AM, George Weaver wrote:
>> Hi List,
>>
>> I need to import data from a large Excel spreadsheet into a PostgreSQL
>> table.  I have a program that uses ODBC to connect to Excel and extract
>> data using SQL queries.  The program then inserts the data into a
>> PostgreSQL table.
>>
>> The challenge with this particular spreadsheet is that it is arrayed
>> thus:
>>
>> Model No 1     Product Code 15
>>     Serial No 1       No on Hand
>>     Serial No 2       No on Hand
>>     Serial No 3       No on Hand
>> Model No 4    Product Code 9
>>     Serial No 12     No on Hand
>> Model No 5    Product Code 27
>>     Serial No 6       No on Hand
>>     Serial No 14     No on Hand
>>
>> etc.
>>
>> I need the data in PostgreSQL arrayed thus
>>
>> Model No 1    Product Code 15    Serial No 1       No on Hand
>> Model No 1    Product Code 15    Serial No 2       No on Hand
>> Model No 1    Product Code 15    Serial No 3       No on Hand
>> Model No 4    Product Code 9      Serial No 12     No on Hand
>> Model No 5    Product Code 27    Serial No 6       No on Hand
>> Model No 5    Product Code 27    Serial No 14     No on Hand
>>
>> I can import the data procedurely using plpgsql to match the individual
>> rows to the master for each row (import the raw data into a buffer table
>> in PostgreSQL and then looping through the rows in the buffer table and
>> checking to see when the Model No changes).
>
> Note that if you're doing that you better be putting the rownumber from
> excel into the table... result sets are NOT guaranteed to be in insert
> order!

Good point.
>
>> I'm wondering if there is a more elegant way to do this using straight
>> sql from Excel?
>
> Well, that's really an excel question, not a Postgres question...
>
> If you load the whole spreadsheet into a single table and have a way to
> differentiate between the different rows then you might be able to do
> something with CTE's to relate a serial number to the product code. That
> might be faster than plpgsql.

I have gone the route of loading the whole file into a buffer table in
PostgreSQL and processing it from there.  Since the only way I can relate
each row to the model number is by its relative row position, I have used a
plpgsql loop to "flatten" the data.

>
> You might also be able to do something creative with formulas in excel to
> copy the product code data to the serial # rows. You could then import the
> whole thing and re-normalize it.

I looked at this but decided to just do it in PostgreSQL - much more
powerful :-)

>
> There's probably some stuff you could do with VBA too. If you care about
> performance you don't want to execute SQL statements for each spreadsheet
> row.
> --
> Jim Nasby, Data Architect, Blue Treble Consulting
> Data in Trouble? Get it in Treble! http://BlueTreble.com