4 ms·

pivot tables on the top of a database That's interesting. As far as I can tell, pivot tables were a hack that MS added to Excel in response to Lotus Improv, wh

by gruseom 15y ago

pivot tables on the top of a database That's interesting. As far as I can tell, pivot tables were a hack that MS added to Excel in response to Lotus Improv, which was an early 90s attempt at a higher-level spreadsheet that Steve Jobs raved about for years. Perhaps you can restore them to their original glory. because excel is really slow with what I call 'small data' and it also has strange row limits Can you elaborate? What do you call 'small data'? And what are the strange row limits? I'm interested in these things because I work in an Excel-related space, which I would be happy to discuss offline (email in profile).

viandante 15y ago
1 = Yes, that's exactly what I think we need here. A perfect example is the "data frame" concept in R (that you can implement as a list of tables in other languages). Excel doesn't have decent high level abstractions of your data, that forces the user to do a lot of boiler plate work that can be extremely time consuming and dangerous (as the more more manual work, the more mistakes). 2 = Excel 2003 limits (still used in my company): http://office.microsoft.com/en-us/excel-help/excel-specifications-and-limits-HP005199291.aspx http://office.microsoft.com/en-us/excel-help/excel-specifica.... Especially when you make deep analysis (first 10 customers of segment B, low margin products, lost due to low margin products, pricing, etc.), you hit the limit and you are forced to do bad things (like using 12 columns instead than 2 to represent 12 months). 3 = As 'small data' I just mean all those data people use in companies. Usually those are the result of extracting data using a business warehouse system (ie cognos) or financial data submitted by entities (ie 30/40 files from different parts of the world of PnLs to consolidate...) or KPIs, or... you get the idea... p.s. I will be sending you an email later today. It would be really great to discuss those concepts.
- gruseom 15y ago
  Please do. I look forward to it.

by gruseom 15y ago

viandante 15y ago

1 = Yes, that's exactly what I think we need here. A perfect example is the "data frame" concept in R (that you can implement as a list of tables in other languages). Excel doesn't have decent high level abstractions of your data, that forces the user to do a lot of boiler plate work that can be extremely time consuming and dangerous (as the more more manual work, the more mistakes). 2 = Excel 2003 limits (still used in my company): http://office.microsoft.com/en-us/excel-help/excel-specifications-and-limits-HP005199291.aspx http://office.microsoft.com/en-us/excel-help/excel-specifica.... Especially when you make deep analysis (first 10 customers of segment B, low margin products, lost due to low margin products, pricing, etc.), you hit the limit and you are forced to do bad things (like using 12 columns instead than 2 to represent 12 months). 3 = As 'small data' I just mean all those data people use in companies. Usually those are the result of extracting data using a business warehouse system (ie cognos) or financial data submitted by entities (ie 30/40 files from different parts of the world of PnLs to consolidate...) or KPIs, or... you get the idea... p.s. I will be sending you an email later today. It would be really great to discuss those concepts.

gruseom 15y ago
Please do. I look forward to it.