Hi, Mark, Got the test case. I am getting seg fault on my test machine. So it seems like there is something wrong, but I am not sure exactly what is the problem yet. Will let you know as soon as I am able to find out more information.
John On 6/12/13 10:57 AM, Mark Hansen wrote: > John, > > Attached are files that demonstrate the problem. > > zip.csv and country.csv each have 3 columns, and 2 in common. So, > when I append them to an (initially empty) partition, I would expect > the result to have 4 columns (the union). However, the -part.txt file > shows only the 3 columns of the last one appended. > > load-commands.txt contains the shell commands used to load the > original CSVs and combine them using append. > > Sorry for the delay getting this to you. > > -- Mark > > > > On Wed, Jun 12, 2013 at 1:47 AM, K. John Wu <[email protected] > <mailto:[email protected]>> wrote: > > Hi, Mark, > > Sorry for the slow response. > > I have not seen an sample data from you yet. Have you resolved the > problem with appending data partitions? > > There are a couple of examples of using ibis -append option in > tests/Makefile.am for targets do-small-append and do-large-append. > Please take a look and see if something similar to that works for you. > > John > > > > On 6/8/13 4:41 AM, Mark Hansen wrote: > > So, I should be able to achieve such a union of columns by > writing some C++ code, right? > > > > I'm building a wrapper to do some specialized loading and queries. > > > > I will send you an example later today. > > > > > > Mark Hansen > > President & Founder > > Digital Brand Mine - developers of Megalytic > > http://megalytic.com > > http://digitalbrandmine.com > > > > On Jun 7, 2013, at 10:40 PM, "K. John Wu" <[email protected] > <mailto:[email protected]>> wrote: > > > >> Hi, Mark, > >> > >> We did intent -a option to create a union of the columns. Note > that > >> this operation will assume the columns did not exist in the first > >> table to be NULLs. > >> > >> Anyway, if you have a small example you can share with us, I > will take > >> a look into this problem over the weekend. > >> > >> John > >> > >> > >> On 6/3/13 8:17 PM, Mark Hansen wrote: > >>> I have 2 tables: X and Y. I want to append the into a new > table Z. > >>> So, I do this: > >>> > >>> ibis -a X to ./Z > >>> > >>> That works fine, and Z is now identical to X. Then, I do this: > >>> > >>> ibis -a Y to ./Z > >>> > >>> It seems to append the rows, but the ./Z/-part.txt file is not the > >>> UNION of X and Y as I expected. Instead, it is identical to > >>> ./Y/-part.txt. The columns from both X and Y are there as > files under > >>> ./Z, it is just the ./Z/-part.txt which does not have all the > columns. > >>> > >>> What am I missing? > >>> > >>> > >>> > >>> > >>> _______________________________________________ > >>> FastBit-users mailing list > >>> [email protected] <mailto:[email protected]> > >>> https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users > >>> > > > > > -- > Mark Hansen > Founder & President > Digital Brand Mine | 708 3rd Ave | New York, New York 10017 > office: 212-961-7250 > http://digitalbrandmine.com/ | email: [email protected] > <mailto:[email protected]> _______________________________________________ FastBit-users mailing list [email protected] https://hpcrdm.lbl.gov/cgi-bin/mailman/listinfo/fastbit-users
