6 ms·

I've had a shocking experience of this product (abysmal performance, memory leaks and poor support). It's basically a poorly wrapped version of pyxll[1] which I

by dia80 13y ago

I've had a shocking experience of this product (abysmal performance, memory leaks and poor support). It's basically a poorly wrapped version of pyxll[1] which I would recommend instead. It's not production software. [1] http://www.pyxll.com/ http://www.pyxll.com/

waterlion 13y ago
I was wondering how IPC was done. I would imagine that VBA works on in-process data structures whereas you'd have to marshal messages over IPC to get Python working. How does it work? Is it more COM-like?
digz 13y ago
Don't think most users are looking to use this in a production environment. Virtually everything you can do with DN and Excel generally, you can do much more effectively without touching excel. It however is extremely useful for one-offs, prototyping, and for those unfortunate situations where you need excel. Additionally, when I've had issues, they've typically been handled promptly and thoroughly.
- waterlion 13y ago
  I'm sorry to say that lots of people are using Excel in 'production' for mission-critical money things. It's scary. For example: http://financialsystems.sungard.com/solutions/hedge-funds-management/Hedge360/convertible-arbitrage/Monis http://financialsystems.sungard.com/solutions/hedge-funds-ma...
karamazov 13y ago
I'm really sorry about that. When did you use the software? We've fixed a number of memory leak issues in the past few months, so this was likely resolved. Additionally, we don't wrap pyxll; we have a completely seperate architecture.
sheetjs 13y ago
It's not trying to target power users. They would learn VBA. It's not trying to target data processing. People who care enough would use a library like xlrd/xlwt. It's not trying to go after big industries like finance. It's easy to interop outside of excel and from within VBA, and most hedge funds and banks have their own solutions already. This is specifically targeted to the casual people who want to do some in-Excel scripting but don't want to learn VBA. In that light, the product seems way over-priced (and probably should be in the $5-10 range) and unlikely to ever be used in a production setting.
- karamazov 13y ago
  I disagree with your points: | It's not trying to target power users. They would learn VBA. Most of our users know VBA and prefer not to use it. (Programmers who've ever tried VBA are familiar with its downsides; I won't go into them here.) Python is a significantly more productive language than VBA; our clients can typically write programs 10x faster in Python than in VBA. More importantly, a large class of programs that are trivial to write in Python using open-source libraries would take months or years of development time in VBA. | It's not trying to target data processing. People who care enough would use a library like xlrd/xlwt. xlrd/xlwt are very good for a subset of what DataNitro does: specifically, static processing (input and output) of Excel files. They don't support more complicated use cases (for example, building an interface to an external data source for non-technical end users, or a spreadsheet that runs analyses based on user input). | It's not trying to go after big industries like finance. We have a number of users in finance. | It's easy to interop outside of excel and from within VBA I wouldn't say it's easy, especially for people who haven't done it before. It'll certainly take more than two hours to figure out, at which point the license has already paid for itself. | most hedge funds and banks have their own solutions already. This is true, but internal solutions are frequently less flexible and harder to use than DataNitro; at the least, this makes them unsuitable for work done by non-developers (e.g. quants). It commonly saves developers time as well. | This is specifically targeted to the casual people who want to do some in-Excel scripting but don't want to learn VBA. Given the option, most people prefer to use Python over VBA; this is true even if (in fact, especially if) they know both languages. | In that light, the product seems way over-priced (and probably should be in the $5-10 range) and unlikely to ever be used in a production setting. DataNitro is used both in production and as a prototyping system by a number of firms.

by dia80 13y ago

waterlion 13y ago

I was wondering how IPC was done. I would imagine that VBA works on in-process data structures whereas you'd have to marshal messages over IPC to get Python working. How does it work? Is it more COM-like?

digz 13y ago

Don't think most users are looking to use this in a production environment. Virtually everything you can do with DN and Excel generally, you can do much more effectively without touching excel. It however is extremely useful for one-offs, prototyping, and for those unfortunate situations where you need excel. Additionally, when I've had issues, they've typically been handled promptly and thoroughly.

waterlion 13y ago
I'm sorry to say that lots of people are using Excel in 'production' for mission-critical money things. It's scary. For example: http://financialsystems.sungard.com/solutions/hedge-funds-management/Hedge360/convertible-arbitrage/Monis http://financialsystems.sungard.com/solutions/hedge-funds-ma...

karamazov 13y ago

I'm really sorry about that. When did you use the software? We've fixed a number of memory leak issues in the past few months, so this was likely resolved. Additionally, we don't wrap pyxll; we have a completely seperate architecture.

sheetjs 13y ago

It's not trying to target power users. They would learn VBA. It's not trying to target data processing. People who care enough would use a library like xlrd/xlwt. It's not trying to go after big industries like finance. It's easy to interop outside of excel and from within VBA, and most hedge funds and banks have their own solutions already. This is specifically targeted to the casual people who want to do some in-Excel scripting but don't want to learn VBA. In that light, the product seems way over-priced (and probably should be in the $5-10 range) and unlikely to ever be used in a production setting.

karamazov 13y ago
I disagree with your points: | It's not trying to target power users. They would learn VBA. Most of our users know VBA and prefer not to use it. (Programmers who've ever tried VBA are familiar with its downsides; I won't go into them here.) Python is a significantly more productive language than VBA; our clients can typically write programs 10x faster in Python than in VBA. More importantly, a large class of programs that are trivial to write in Python using open-source libraries would take months or years of development time in VBA. | It's not trying to target data processing. People who care enough would use a library like xlrd/xlwt. xlrd/xlwt are very good for a subset of what DataNitro does: specifically, static processing (input and output) of Excel files. They don't support more complicated use cases (for example, building an interface to an external data source for non-technical end users, or a spreadsheet that runs analyses based on user input). | It's not trying to go after big industries like finance. We have a number of users in finance. | It's easy to interop outside of excel and from within VBA I wouldn't say it's easy, especially for people who haven't done it before. It'll certainly take more than two hours to figure out, at which point the license has already paid for itself. | most hedge funds and banks have their own solutions already. This is true, but internal solutions are frequently less flexible and harder to use than DataNitro; at the least, this makes them unsuitable for work done by non-developers (e.g. quants). It commonly saves developers time as well. | This is specifically targeted to the casual people who want to do some in-Excel scripting but don't want to learn VBA. Given the option, most people prefer to use Python over VBA; this is true even if (in fact, especially if) they know both languages. | In that light, the product seems way over-priced (and probably should be in the $5-10 range) and unlikely to ever be used in a production setting. DataNitro is used both in production and as a prototyping system by a number of firms.