Hierarchical data structure from a multi-dimensional array
in [TCL]
|
From: yilmaz on 18 Jul 2010 03:08 Hi all, I would appreciate any help. I have a multi-dimensional array representing the column and row dimensions in a table and values in the table cells: --0---------1-------------2-------3-----------4-------------- FruitA,category1,type1,name1 0.1345 FruitA,category1,type2,name1 0.1123 FruitA,category2,type1,name1 0.1993 ............................... FruitA,category2,type2,name2 0.1183 ................................ FruitA,category3,type1,name2 0.1100 ......... FruitB,category1,type1,name1 0.0100 FruitB,category1,type2,name1 0.1123 ........ FruitC,category1,type1,name1 0.1403 .......................... In the array above, FruitA,category1,type1 represents the column dimensions in the table, name1 represents a row dimension, and the 0.1345 represents a value in the given cell in the table. I need to transfer the array dimensions above into a hierarchical flat list to first configure my column and row headings and sort them alphabetically before dumping these into an XML-like tree. My list will be constructed such as to have one-to-one correspondence between the related dimensions like this: set columnHeadingsList { {FruitA FruitA ... FruitB FruitB ... FruitC ...} # sorted alphabetically {{category1 category2 ....} {category1 category2 ...} } #sorted alphabetically {{type1 type2 ...} {type1 type2 ...} {type1 type2 ...} {type1 type2 ...}} # sorted alphabetically } set rowHeadingsList { {name1 name2 ....} # sorted alphabetically } I have the methods to extract the dimensions to form the lists and can do the sort alphabetically. My concern is the following, though. When I set up the column headings and row headings in these list, I will have to implement some recursive tree traversal method that will remember the path to the parent nodes to eventually extract the whole path with dimensions. For example, when I give the row and column index , it should return FruitA,category1,type1,name1 and then using some mapping method I can verify if the entry with such index exists in the original array and if it exists the value 0.1345 for the key FruitA,category1,type1,name1 should be extracted and put into the corresponding XML tag. I looked at the tree traversal methods but found nothing that satisfies my requirements. Has someone maybe more experience in table list traversal methods that would suit my case ? I'd appreciate any help with this. Thank you. YY.
From: Bezoar on 18 Jul 2010 04:04 On Jul 18, 2:08 am, yilmaz <yyyyil...(a)gmail.com> wrote: > Hi all, > > I would appreciate any help. > > I have a multi-dimensional array representing the column and row > dimensions in a table and values in the table cells: > > --0---------1-------------2-------3-----------4-------------- > FruitA,category1,type1,name1 0.1345 > FruitA,category1,type2,name1 0.1123 > FruitA,category2,type1,name1 0.1993 > .............................. > FruitA,category2,type2,name2 0.1183 > ............................... > FruitA,category3,type1,name2 0.1100 > ........ > FruitB,category1,type1,name1 0.0100 > FruitB,category1,type2,name1 0.1123 > ....... > FruitC,category1,type1,name1 0.1403 > ......................... > > In the array above, FruitA,category1,type1 represents the column > dimensions in the table, name1 represents a row dimension, and the > 0.1345 represents a value in the given cell in the table. > > I need to transfer the array dimensions above into a hierarchical flat > list to first configure my column and row headings and sort them > alphabetically before dumping these into an XML-like tree. > > My list will be constructed such as to have one-to-one correspondence > between the related dimensions like this: > > set columnHeadingsList { > {FruitA FruitA ... FruitB FruitB ... FruitC ...} # sorted > alphabetically > {{category1 category2 ....} {category1 category2 ...} } #sorted > alphabetically > {{type1 type2 ...} {type1 type2 ...} {type1 type2 ...} {type1 > type2 ...}} # sorted alphabetically > > } > > set rowHeadingsList { > {name1 name2 ....} # sorted alphabetically > > } > > I have the methods to extract the dimensions to form the lists and can > do the sort alphabetically. > > My concern is the following, though. When I set up the column headings > and row headings in these list, I will have to implement some > recursive tree traversal method that will remember the path to the > parent nodes to eventually extract the whole path with dimensions. > > For example, when I give the row and column index , it should return > FruitA,category1,type1,name1 and then using some mapping method I can > verify if the entry with such index exists in the original array and > if it exists the value 0.1345 for the key > FruitA,category1,type1,name1 should be extracted and put into the > corresponding XML tag. > > I looked at the tree traversal methods but found nothing that > satisfies my requirements. Has someone maybe more experience in table > list traversal methods that would suit my case ? I'd appreciate any > help with this. > > Thank you. > YY. If I understand you correctly you have an array with your key being the column/row as index and the value being some number. From the array you have extracted the columns and row indexes and have alphabetized them. Now you want to regurgitate the array in xml but do it alphabetically by index. Your method would seem to require foreach fruit [lindex $ColumnHeadingList 0 ] { foreach category [lindex $ColumnHeadingList 1 ] { foreach type [lindex $ColumnHeadingList 2 ] { foreach row $rowHeadingsList { if { [ info exists dataArray($fruit,$category, $type,$row) ] } { outputToXML "$fruit,$category,$type,$row" \ $dataArray($fruit,$category,$type,$row) } } } } However it may be easier to pad the numbers of the columns and use the dictionary sort so convert your column1 to column00001 and type1 to type00001 etch for all column types and rows. This way if your number portion exceeds 9 then 00001 - 00009 will sort before 00010 and higher. Then to quickly sort all your data and output would be this. foreach idx [ lsort -dictionary [array names dataArray ] ] { outputToXML [removePadding $idx ] $dataArray($idx) } Bezoar Note: I just typed these in so minor syntax errors may have crept in
From: Goga Yatzenko on 18 Jul 2010 06:49 On Jul 18, 1:04 pm, Bezoar <cwjo...(a)gmail.com> wrote: > On Jul 18, 2:08 am, yilmaz <yyyyil...(a)gmail.com> wrote: > > > > > Hi all, > > > I would appreciate any help. > > > I have a multi-dimensional array representing the column and row > > dimensions in a table and values in the table cells: > > > --0---------1-------------2-------3-----------4-------------- > > FruitA,category1,type1,name1 0.1345 > > FruitA,category1,type2,name1 0.1123 > > FruitA,category2,type1,name1 0.1993 > > .............................. > > FruitA,category2,type2,name2 0.1183 > > ............................... > > FruitA,category3,type1,name2 0.1100 > > ........ > > FruitB,category1,type1,name1 0.0100 > > FruitB,category1,type2,name1 0.1123 > > ....... > > FruitC,category1,type1,name1 0.1403 > > ......................... > > > In the array above, FruitA,category1,type1 represents the column > > dimensions in the table, name1 represents a row dimension, and the > > 0.1345 represents a value in the given cell in the table. > > > I need to transfer the array dimensions above into a hierarchical flat > > list to first configure my column and row headings and sort them > > alphabetically before dumping these into an XML-like tree. > > > My list will be constructed such as to have one-to-one correspondence > > between the related dimensions like this: > > > set columnHeadingsList { > > {FruitA FruitA ... FruitB FruitB ... FruitC ...} # sorted > > alphabetically > > {{category1 category2 ....} {category1 category2 ...} } #sorted > > alphabetically > > {{type1 type2 ...} {type1 type2 ...} {type1 type2 ...} {type1 > > type2 ...}} # sorted alphabetically > > > } > > > set rowHeadingsList { > > {name1 name2 ....} # sorted alphabetically > > > } > > > I have the methods to extract the dimensions to form the lists and can > > do the sort alphabetically. > > > My concern is the following, though. When I set up the column headings > > and row headings in these list, I will have to implement some > > recursive tree traversal method that will remember the path to the > > parent nodes to eventually extract the whole path with dimensions. > > > For example, when I give the row and column index , it should return > > FruitA,category1,type1,name1 and then using some mapping method I can > > verify if the entry with such index exists inthe originalarray and > > if it exists the value 0.1345 for the key > > FruitA,category1,type1,name1 should be extracted and put into the > > corresponding XMLtag. > > > I looked at the tree traversal methods but found nothing that > > satisfies my requirements. Has someone maybe more experience in table > > list traversal methods that would suit my case ? I'd appreciate any > > help with this. > > > Thank you. > > YY. > > If I understand you correctly you have an array with your key being > the column/row as index and the value being some number. From the > array you have extracted the columns and row indexes and have > alphabetized them. Now you want to regurgitate the array in xml but do > it alphabetically by index. Your method would seem to require > foreach fruit [lindex $ColumnHeadingList 0 ] { > foreach category [lindex $ColumnHeadingList 1 ] { > foreach type [lindex $ColumnHeadingList 2 ] { > foreach row $rowHeadingsList { > if { [ info exists dataArray($fruit,$category, > $type,$row) ] } { > outputToXML "$fruit,$category,$type,$row" > \ $dataArray($fruit,$category,$type,$row) > } > } > } > > } > > However it may be easier to pad the numbers of the columns and use the > dictionary sort > so convert your column1 to column00001 and type1 to type00001 etch for > all column types and > rows. This way if your number portion exceeds 9 then 00001 - 00009 > will sort before 00010 and higher. Then to quickly sort all your data > and output would be this. > > foreach idx [ lsort -dictionary [array names dataArray ] ] { > outputToXML [removePadding $idx ] $dataArray($idx) > > } > > Bezoar > > Note: I just typed these in so minor syntax errors may have crept in Thank you for your reply and the pointers, Bezoar! They help a great deal. Besides the printing of column headings and rows, I was also looking for a way to extract the values given the key. Like if I have extracted FruitA,category1,type1,name1 from the list , then I would go and map the FruitA, category1, type1, and name1 to form the right index in the original array to grab the value for that key. So, for example, if I have FruitA,category1,type1,name1 key from the list, then there will be a mapping method that would set the dimensions like: set columns { {fruit 0} {category 1} {type 2} } set rows { {name 4} } Then, the key FruitA,category1,type1,name1 would be set through these methods to give the real positions of items in the original array. If the given key after mapping is found in the array, then its value is extracted and put into the table cell. At that time, all column headings and row headings will already be printed into XML. Only the values like 0.1345 will be printed at the end when the table headings are in place. --0-------1----------2-------3----4-------------- FruitA,category1,type1,name1 0.1345 Maybe I am doing too many checks, but that way I guess I will have the correct values printed into the table cells. That's why I was thinking about a recursion through a tree to get the key bindings from the list . Thank you for your help. If you can just comment if I am on the right track with this implementation, I'd be very grateful.
From: yilmaz on 18 Jul 2010 06:51 Thank you for your reply and the pointers, Bezoar! They help a great deal. Besides the printing of column headings and rows, I was also looking for a way to extract the values given the key. Like if I have extracted FruitA,category1,type1,name1 from the list , then I would go and map the FruitA, category1, type1, and name1 to form the right index in the original array to grab the value for that key. So, for example, if I have FruitA,category1,type1,name1 key from the list, then there will be a mapping method that would set the dimensions like: set columns { {fruit 0} {category 1} {type 2} } set rows { {name 4} } Then, the key FruitA,category1,type1,name1 would be set through these methods to give the real positions of items in the original array. If the given key after mapping is found in the array, then its value is extracted and put into the table cell. At that time, all column headings and row headings will already be printed into XML. Only the values like 0.1345 will be printed at the end when the table headings are in place. --0-------1----------2-------3----4-------------- FruitA,category1,type1,name1 0.1345 Maybe I am doing too many checks, but that way I guess I will have the correct values printed into the table cells. That's why I was thinking about a recursion through a tree to get the key bindings from the list . Thank you for your help. If you can just comment if I am on the right track with this implementation, I'd be very grateful. On Jul 18, 1:04 pm, Bezoar <cwjo...(a)gmail.com> wrote: > On Jul 18, 2:08 am, yilmaz <yyyyil...(a)gmail.com> wrote: > > > > > Hi all, > > > I would appreciate any help. > > > I have a multi-dimensional array representing the column and row > > dimensions in a table and values in the table cells: > > > --0---------1-------------2-------3-----------4-------------- > > FruitA,category1,type1,name1 0.1345 > > FruitA,category1,type2,name1 0.1123 > > FruitA,category2,type1,name1 0.1993 > > .............................. > > FruitA,category2,type2,name2 0.1183 > > ............................... > > FruitA,category3,type1,name2 0.1100 > > ........ > > FruitB,category1,type1,name1 0.0100 > > FruitB,category1,type2,name1 0.1123 > > ....... > > FruitC,category1,type1,name1 0.1403 > > ......................... > > > In the array above, FruitA,category1,type1 represents the column > > dimensions in the table, name1 represents a row dimension, and the > > 0.1345 represents a value in the given cell in the table. > > > I need to transfer the array dimensions above into a hierarchical flat > > list to first configure my column and row headings and sort them > > alphabetically before dumping these into an XML-like tree. > > > My list will be constructed such as to have one-to-one correspondence > > between the related dimensions like this: > > > set columnHeadingsList { > > {FruitA FruitA ... FruitB FruitB ... FruitC ...} # sorted > > alphabetically > > {{category1 category2 ....} {category1 category2 ...} } #sorted > > alphabetically > > {{type1 type2 ...} {type1 type2 ...} {type1 type2 ...} {type1 > > type2 ...}} # sorted alphabetically > > > } > > > set rowHeadingsList { > > {name1 name2 ....} # sorted alphabetically > > > } > > > I have the methods to extract the dimensions to form the lists and can > > do the sort alphabetically. > > > My concern is the following, though. When I set up the column headings > > and row headings in these list, I will have to implement some > > recursive tree traversal method that will remember the path to the > > parent nodes to eventually extract the whole path with dimensions. > > > For example, when I give the row and column index , it should return > > FruitA,category1,type1,name1 and then using some mapping method I can > > verify if the entry with such index exists inthe originalarray and > > if it exists the value 0.1345 for the key > > FruitA,category1,type1,name1 should be extracted and put into the > > corresponding XMLtag. > > > I looked at the tree traversal methods but found nothing that > > satisfies my requirements. Has someone maybe more experience in table > > list traversal methods that would suit my case ? I'd appreciate any > > help with this. > > > Thank you. > > YY. > > If I understand you correctly you have an array with your key being > the column/row as index and the value being some number. From the > array you have extracted the columns and row indexes and have > alphabetized them. Now you want to regurgitate the array in xml but do > it alphabetically by index. Your method would seem to require > foreach fruit [lindex $ColumnHeadingList 0 ] { > foreach category [lindex $ColumnHeadingList 1 ] { > foreach type [lindex $ColumnHeadingList 2 ] { > foreach row $rowHeadingsList { > if { [ info exists dataArray($fruit,$category, > $type,$row) ] } { > outputToXML "$fruit,$category,$type,$row" > \ $dataArray($fruit,$category,$type,$row) > } > } > } > > } > > However it may be easier to pad the numbers of the columns and use the > dictionary sort > so convert your column1 to column00001 and type1 to type00001 etch for > all column types and > rows. This way if your number portion exceeds 9 then 00001 - 00009 > will sort before 00010 and higher. Then to quickly sort all your data > and output would be this. > > foreach idx [ lsort -dictionary [array names dataArray ] ] { > outputToXML [removePadding $idx ] $dataArray($idx) > > } > > Bezoar > > Note: I just typed these in so minor syntax errors may have crept in
From: tom.rmadilo on 18 Jul 2010 15:01 On Jul 18, 3:51 am, yilmaz <yyyyil...(a)gmail.com> wrote: > Thank you for your reply and the pointers, Bezoar! They help a great > deal. > > Besides the printing of column headings and rows, I was also looking > for a way to extract the values given the key. Like if I have > extracted FruitA,category1,type1,name1 from the list , then I would > go and map the FruitA, category1, type1, and name1 to form the right > index in the original array to grab the value for that key. > > So, for example, if I have FruitA,category1,type1,name1 key from the > list, then there will be a mapping method that would set the > dimensions like: > > set columns { > {fruit 0} > {category 1} > {type 2} > > } > > set rows { > {name 4} > > } > > Then, the key FruitA,category1,type1,name1 would be set through these > methods to give the real positions of items in the original array. > > If the given key after mapping is found in the array, then its value > is extracted and put into the table cell. At that time, all column > headings and row headings will already be printed into XML. Only the > values like 0.1345 will be printed at the end when the table headings > are in place. > > --0-------1----------2-------3----4-------------- > FruitA,category1,type1,name1 0.1345 > > Maybe I am doing too many checks, but that way I guess I will have the > correct values printed into the table cells. That's why I was thinking > about a recursion through a tree to get the key bindings from the > list . > > Thank you for your help. If you can just comment if I am on the right > track with this implementation, I'd be very grateful. > > On Jul 18, 1:04 pm, Bezoar <cwjo...(a)gmail.com> wrote: > > > > > On Jul 18, 2:08 am, yilmaz <yyyyil...(a)gmail.com> wrote: > > > > Hi all, > > > > I would appreciate any help. > > > > I have a multi-dimensional array representing the column and row > > > dimensions in a table and values in the table cells: > > > > --0---------1-------------2-------3-----------4-------------- > > > FruitA,category1,type1,name1 0.1345 > > > FruitA,category1,type2,name1 0.1123 > > > FruitA,category2,type1,name1 0.1993 > > > .............................. > > > FruitA,category2,type2,name2 0.1183 > > > ............................... > > > FruitA,category3,type1,name2 0.1100 > > > ........ > > > FruitB,category1,type1,name1 0.0100 > > > FruitB,category1,type2,name1 0.1123 > > > ....... > > > FruitC,category1,type1,name1 0.1403 > > > ......................... > > > > In the array above, FruitA,category1,type1 represents the column > > > dimensions in the table, name1 represents a row dimension, and the > > > 0.1345 represents a value in the given cell in the table. > > > > I need to transfer the array dimensions above into a hierarchical flat > > > list to first configure my column and row headings and sort them > > > alphabetically before dumping these into an XML-like tree. > > > > My list will be constructed such as to have one-to-one correspondence > > > between the related dimensions like this: > > > > set columnHeadingsList { > > > {FruitA FruitA ... FruitB FruitB ... FruitC ...} # sorted > > > alphabetically > > > {{category1 category2 ....} {category1 category2 ...} } #sorted > > > alphabetically > > > {{type1 type2 ...} {type1 type2 ...} {type1 type2 ...} {type1 > > > type2 ...}} # sorted alphabetically > > > > } > > > > set rowHeadingsList { > > > {name1 name2 ....} # sorted alphabetically > > > > } > > > > I have the methods to extract the dimensions to form the lists and can > > > do the sort alphabetically. > > > > My concern is the following, though. When I set up the column headings > > > and row headings in these list, I will have to implement some > > > recursive tree traversal method that will remember the path to the > > > parent nodes to eventually extract the whole path with dimensions. > > > > For example, when I give the row and column index , it should return > > > FruitA,category1,type1,name1 and then using some mapping method I can > > > verify if the entry with such index exists inthe originalarray and > > > if it exists the value 0.1345 for the key > > > FruitA,category1,type1,name1 should be extracted and put into the > > > corresponding XMLtag. > > > > I looked at the tree traversal methods but found nothing that > > > satisfies my requirements. Has someone maybe more experience in table > > > list traversal methods that would suit my case ? I'd appreciate any > > > help with this. > > > > Thank you. > > > YY. > > > If I understand you correctly you have an array with your key being > > the column/row as index and the value being some number. From the > > array you have extracted the columns and row indexes and have > > alphabetized them. Now you want to regurgitate the array in xml but do > > it alphabetically by index. Your method would seem to require > > foreach fruit [lindex $ColumnHeadingList 0 ] { > > foreach category [lindex $ColumnHeadingList 1 ] { > > foreach type [lindex $ColumnHeadingList 2 ] { > > foreach row $rowHeadingsList { > > if { [ info exists dataArray($fruit,$category, > > $type,$row) ] } { > > outputToXML "$fruit,$category,$type,$row" > > \ $dataArray($fruit,$category,$type,$row) > > } > > } > > } > > > } > > > However it may be easier to pad the numbers of the columns and use the > > dictionary sort > > so convert your column1 to column00001 and type1 to type00001 etch for > > all column types and > > rows. This way if your number portion exceeds 9 then 00001 - 00009 > > will sort before 00010 and higher. Then to quickly sort all your data > > and output would be this. > > > foreach idx [ lsort -dictionary [array names dataArray ] ] { > > outputToXML [removePadding $idx ] $dataArray($idx) > > > } > > > Bezoar > > > Note: I just typed these in so minor syntax errors may have crept in It is hard to figure out what hierarchy you are trying to get at. What does one XML record look like? However, I have two suggestions for modeling multi-dimensional arrays with Tcl arrays: 1: Use a formal Tcl list as the key: set md_array([list $fruit $cat $name]) $value 2. Sort by using multiple use of [lsort]: set keys [array names md_array] set sorted_keys_2_1_0 [lsort -index 0 [lsort -index 1 [lsort -index 2 $keys]]] Look at the [lsort] manpages to see if additional options would help you out. The reason to use a Tcl list as a key is to make sure quoting is done properly and to aid in sorting based upon the list elements.
|
Next
|
Last
Pages: 1 2 Prev: Compiling tcl 8.5.8 windows x64 Visual Studio 2008 Next: Font compatibility problem or bug |