<\body> What did we learn last time? <\itemize-dot> matrices: create with . Address entries in an array: 8]> Apply a function on all rows, or all column of an array:\ <\itemize-dot> - calculate mean of all rows. - calculate mean of all columns. All this is very useful for bootstraping! Functions: Read data from file: \; We want to read the following table: L\\Z\\ \\Z$dbtdTVtr4\|zLJ\\\d $,.,LN\\LNL\<\>\<\\4DFTdf,\|z̔42T\\^\|~\\D46tv,$ĄTVTTV$dfd d\\l ,,.4LN$dD-@(Ԓ`!@Pft\0=\0@\04\0\0\0\0\0\0\0\0\0\0-\0\0@\0$D\0\0(\0\0\0\0\0\0\0P\0qt\0!\0@(=@`\|D@@`\0\0\0\0\0\0Pt \0\0\0,\0\0\0\04\0\0\0)Hp 80ÆJHqņ /jqǁ\0ĉ#Ȓ(O42e˕.cœL4oϝ?{HѢH̘R\0JBuJ5զSj@O` J6lٱfӢ]ksַ\\^nBV+7޾d9ƒ+68q‹Kd˔/#͟#wVeh\|Mw+]\0 +l\0d-Ыfѡ+'cG#NWhb;ߵk j dH^y 70dKn\\%Z\|_bBXE{Eg{gJ+QB].@yEH(x$4Tc6/2c:\ni$CX$G*ɣ(]7bW\nqkeaV^ Dm)xcfa04Xp)snFvIZak駛zgxZtTRSWeYxNG9/2$(Z"Ҫʪ~*A#OJkkjXiZ(]iʲXC`l+j€])S꒯nP$kMkO9e$ae[nT6Б%to+k\t\0-.mqa2[0RHP_*N)LJ]upQVZ93\>(+(Guӫ*JY1sMkWn-voXv!KtbJZXq붽b c Q-Ln'7G.9CTngw砇.褗n:瓧ꬷx^./ o_\|7G/Ooc/' w݃NA~ODvv7@w"H\>\0 w';ڮyd_@\nf aA\tZ s6;\nKH.L \t/k]ArXPD`CP,'j\0AD=P~S,(\<,_LE4Zόx%\0Ѐ (8\0 PE\0HAA\0X HBDBЋ_Т0`Bd%/ePD \01j))\<R0dp_b$%y;?6 f^ς,qg`QTb\0 N\n\\x@0dE.PܡS`6ٯP)Z2%)C\|o\nQTX\0qvBሠD@Pp'\\H6J\0R$%\\-sQ}tF5\nQ 3@%T\<$eED'RA;7f.E)Vmħ LO)Ԕ5XW9.zH\n[LbOe+\>X\0aaՀ&T0v\00 I\0\00,\0#uzeEd'[B \0\0Iآ\t-gZ7Vwe1YVh 'ؽlSHQh \0VvX 5\0 $\|\0v\0@0(;Pb jz׷ o"S-k'1^0m򴈐mb`eR"jo\0l2n vw(H3L\0U""pW b%\>PZJ^P;+w&F1~7̊ì\\:y1+ZZy½\0\0J\0\|vO}(\0\05̹#e5`r 8UYȓcA(Q]H"À$\te~LtPmV!Z*i͸t, \0Ikg\<tme\>շe3jCgXʻFZ\0ŤkfujP כ` +t~tE¹:R p5d7Ң.1/mt=͊z{޿{eML-8u?mO[TxgH%f&VÐ\n\>ĭa[\t$S3Ӷy7@o5_ydN\00IJT+]LxYpF972BqjyZ}~\`oVX-}#`}*cz\0χ;]aztD[\t \0\0MPk^ff`("hT~Wx0kVZhIv`t\0_zP0xpIֆoshȇ \0\>pԊ=X\<Z׳z恋=RR w^sƌkd\<(xڣg3\>DP8PrRX?8uӍȊ;x\<\0%\t߸yX8=䑏ӊ(XcE9\nEy%[ّ  "Y&9$yVAD;\0` 0)196Y8I:y;\<\>=9@)O\nJLٔNPR9TITyXZ\\ٕ^`9VdYfcye?\t1GFp9oYrityz\|~i iɕYy阐9ZԒ}\t2IpI9O\0Д`8LQyٛY)ly\nQiQ\t\0yٙ\t\ \0\0D[\tJa!\t^`\tU\0\t"+`\t\0W6 0[P(`\t\0\0 +rP\np\tƋ \0kAꝿS \0jPFP\nt\0_ +5QP Ǜ *: B. n![i JF\0p&F ꕧ2 : 鸎;9T\n[MyET`\td" ` JM 01"[\tJJY\n&֩G K\n=;"J)?̔\>KY\0[\0J\0D\njJQLi\|V30QO@{wl1y+0 : ,ɑ{[JK\00 L7kOɪ1욢LI \0ܰܰSL"a ^"{J)\0˿9:ò8z˱,Ex\\ \<' OZ8\|8~Y Jp3L \\'\0NL) ]лlM1:R\n\0\0,[9\n0\0\0S0MYxy\\Lǒ* 19 \0PǞ?Ivy顗l\>ë3̜˪)EP։̚1Ϭ мOյl=\t)B,}Lɴ٩ \ \t \0 ց\t+̜2Y\0o){֥G ׅ飂`OƯ\0\0S\0J)wШK\0W͛=ָ;\nt\0\0\>`\|\0\n\00/p9Z#W\<)-5 Ơ yVn@ \0GVގ;A}Cͧy \nU@T\nJ)[p1׫5Tp\t˺˸J9~ڱ­)@ \06սtp'@\tKIBla7[1ƀؔ{lӿ3]8p 0\tɗM\>*9ꔥ`Z7p8%K)9@!n#ްmn8o^XM-"b\nN8LД[K^b=5Ij,]r.`\0\0`0 oӘ\>\>ܔ'\0\n{~JI3@2\t =\tE۷^칁ʎ\|e[\>iM0\nFX!\|ߓܗ6@vaΫ?Z`q\0\>\<Ҍk k$O#_Z䑡-,/wz (^\>NvZ/6oE%?O-A_UigIfI2J./\>XB?_,5u_xOGj\|Dw{@y,*\\?_.0OO_pY/_d=O/M_?R/ܶOA(_Cge 2TO)Ь=p\nMɩ6pi@ԹL9@ [ jI\0J@ DPB) \0 aĆ )VĘQP\mP%Ї+H\0BX\\տw@W/`/ƹ=\\\n\t\0E9~A:6)\\.*-HJ1҈+"\n0ˣL5%^z\0 \02dL&\|V(A$.F\01` SLƌ!X'bXP9乑Bḗ:\09! =rp 8AȠ#_chU$.?P\0xD~rl cE\n "d\\TbVZ{1M*bփk9 9a2N+ UĂOf\08APC:AcV5N uy\|+Q蹞%\\шqJk&Y%.Z7,i.PlrE,`q\0MYK—j#t8S򋠄,:cZn($8 0ĩrDr3V#'U\r/wD=8\>\0p"'I\0Q~\t\\qXU^a\nT\07 btvԣȁT]6nW\0C\>R1KCj( EO\0T֙I4P'\0X垜1[ZmlǓ(\\^Jn e*"pe[\n7Ftl@p,jD\>E0Z2ǸLG-9p2ZĴR$Yƫ' H\0=mM=[$O7]V\\\\\0\>9 0\0xr[ Oz:df}\\RpmZ*~5\0ؕJIxX#LTɩ{U嚾AL؂U-F??:h-HP I\tƐ@5W\<֒8G?\tO@,x{=;ē-2?S?h"#BX#7\nHBBYB%\>P²2&ܲ$+{D,Cw\tà[vY(@7\>{m61̈2MbH(7%\0C/Dl-+t:45r6̶9l K:B;T\itC+ť8[J4B ʫ ʞ 8ɯ*8 `@\0F9ԓ'EԲCF%ȰXx\td \08A\<ȁP/”tDd\t`H92 9(f(ElD ̍\nT3tL A \\D\<,4ܔ\t4IypdJL\n*k+ӄKKK(_1` 1Jܪ C*\Ѭ\0ؘWl+X\\ʨtMy\nS)E}YI\<:L}ծV\0.RZ{ɴCj&H]]ͷ#@#:$J4gJݻғN *1R' 6\tu2!-P-  \ Y5:)kJR9!E(A1]{鰛,\0FFRr1J-zҚ"TGP#/=-@ͧPa,\\Vo&8\tFT [,0K0΋\0-W,8tۉϋ,Š%^?Q $afۂbZK\0:2B*j :\>UK\\b\08u1ݻ\\jp?\|+&\> RS)b \|H)#0"2\n DEЅ܉4\nТUќ'@_jaE6^2I"=QQW\0cdͰ\t \|b_AC\|ݕ惨Qc3R^NU.3&gB㱓hc.'XՉjҁؘ`8\nc^%FU"s7M\0™H1&B`C8*C@Bbbz+itFȪ-$*9v9fd!h{P\tܜ5WLBaX\n_%AP?av)DR0LMk~*gvg^S P#AK뺰ku⃅xxkы\t\|nE#\n9N#B*B7\>b7caF$ϸ\nn(PxۂiN`9\<@J@(\0 $Nd8j?&UZִ\0U CBȖtm VPa^llh.kM\$DB\\#\0m @P#Nx7wwM#4yTpƇ*-uHzJw ov@//"3 QxAFȮyQOeO%zhPUaBx`(^.@`67xd@\tF/ E;)D益ynjY$[?Z`!w^g4C ~oam" B艱V ~ELJ\|8@&E S:R҇ 9~2VVc& 8N-\07p+č\010fC1:\0${b „\n\>1C'B4Hƌ7rcF\<\0Fu\0`㖓ix$̘2gҬi&Μ:ateW\tKd\0$C\0)UN"G\t\n\0Q ҝbcQ$IWL5JO(̝K.Lh9[.`\0\>46\0@䂇b ?1YAd7W!a[i@$a3fUҲgӮm[lOF12\0\0j*F54\0bjZWbPMo]tj9#qu\02Z[\\9E*l.\0R6G\0CDƘ-5h\tBfuG\0\\QFF!!ȓka\02\0\nQ&T'hXdk#aiHVؐ_e) e hgY2 QԈ\0\0* M\n\0.\0Q&q9[n9RX)eцxI{9L\<5f0myH(Q{Ub9 \0\>_jj8ٖ'XY\0t)\0F@&kGu+F8*"v+m0kL','f-QBi]UBCajbI\0ePI*VrPBLpFuB ;\VC1#޺IxMQ)6hC5Gnި[ t`.}*M&zI-Ⅹ\t5" (?l= Ұfmd5Rdi0rxb \04"p.Ub9H{a e:$4l0D(a\tB0גFTI)ήT̉?1{.Db\>} 2 ,a -rs[KLL\t\>714K!Df\\\t#fy4c '/E2R'4G}AO#PV2\|+[Sde+]⒗,e)`.x9&2e2t_)iRּ&6m.3&8ÙLos%:ψXJ\nB,ѳ'\>ωcmd(=˒Dx. S?Ġ@3?arQ'HOh:PJ1~t(LzW-d\\i\n\<d2wY*($9@`@aDwb*Ԫnlp@W\nn,?ݷ5C]BKfRXDR9a{FjĪ9ZHvYU ag`[G0[֐R$LC 0H"Q]]t5%T RRX+:"B&kTQu%,aW0@Fp8A^G]qd35G\>\0 *چPt ^uh\0`+ 7a[֪+#%\0$+)J#= {h#@J}j C7s9f d !D ,,1m"&9]hX\|@0aM IN_D2dP\t9-lA2,ט\tl@B*ʁeT mA0\n6B\t)!rgPQE@DH~ Z(pT d0R0!7F"=JNQn9V+gy=3~z\0\0:㵯lWxj"LJ \0`C[&d Xx2ז-ru"E#ߜI#,+&W0\0+[@\\$Ș47 -kA`$%"-+\nu*5\t9f\0#W0`lDF\>ޖr*;#XiKL`x ]ZI']K(='t ݤ}\n,\n:#\0*@I3;W\\t7P,1yuq;\0' tGVަ`\|*UY_=45eMd5o\|{8S\0`ϻLEDgjp)\0\a89 %X0`Ft(qxUk$r XBiĿ`@n\0N-GKe_LF\\@V F˿PΠF `Րk8 *Ufs\ntpUTQauqDǐ\\(:LBJԋCy=}GD/4QU 4[/`D \0Hd\t"aFD#0AkHFD eIB$v ꐁmM&nkTb݉M@+j,D"!ލ"2#fı@j!ya,!=\tpxL%,\ndqW\`Xc\tkt^q_D2Vd\|y@@ga7[}@@yݘaل\nOa\n\%EU^eVne@r OB21G D/@4EP2\\XCt\0\0CC }1`# bF\|\0 ]k …К+ 4sdDoX\\"߰\0FX\\M@%&$ډsd\\l'"&&ogt`V&utmD#If\0#\0@!\01\0 \0xt\0h8U EÌ\\\0D\0B"q\\Iڬ QP FE X#IG+f'a\\\0U޹AX!yX+hUe- \ TN}ZXTf)ڬ݄h hQP%8'\0@!&2J.= MB6B[D`%fX9"ܢI)hsT6%\0ϻygŦ\\]8+ d(QEEaVq\nZTNNk kj[MԎjNBE\<j\|bڑiUN0\0U\t)*2lO`5(*MDoj O@-rj)\0\0\tDl\0)@P`B%@$DP"JdbD%`Q\0% \0&'PlA4mfvn&B$"yl"*,PʀB\|BʁHkni+J\>NSjI\<֒{=Iʄl**tFXr"ѭNMtKjDBWDƫn!F`mpEB\tD@│X\0ȁs@qn: -}m\n؊#'I*lЮR&.YL.QK\n%QZJB'l AIt¥nrnz&o/H .Mf)/[x-R0/-0R/ /+40GO_0[0?r4BS+T\t\t[/ c sD9p$y\t\t\t.\\\|& E\|ppJ0xj_z)QKgYirPƃiSSm#-u-ւ]Qelm]Yx~q夤 IRY؅--hFg _m+\\mۥF1'AYɱ/\\ q~d'GWE "\0'rLKn Jx&[B&$rBIVs)ll@#\\)VlZG85q!B\O؃[c\> 6G\\~,;\\x&TDOAzu\|C\|xhprf`2L1\0{c\nL@R\n\00e`U:( }+\A\0ncO`:\0{\|qgϛmN\0R\0LyX,i@'Yi\05\0KG\>M1\0B7j\0\\p\n+(\\enQ\0PoѸp\0KǗ"7.ys\t]I ;VM?,P -0't\nck!\0IDc\n7 \0t˘3\\\0 BRI;ANs󪠂F\0`\n\0N8!-\0 7\t3 rd\n\n^\0p%\0@"Z\0!4d8/PL+\tN9ɸ-)*QIO#\n%JzHŜ7Q´PRI=\0a5K(\0A*rK\n1l6ò6Qp'kCcSA043XddJ\0 Luת,rࡤ!ܦWf}HaP\>bpDD F"\tJa2WY'SVHT*!Jye\tԕ"} \nm'$C#`13x7`z3B]h`1\0^%4(]9O!:Pd"KfSL*)Tr9x(O`ANiWPdU\>_衑o\\#-bh"M3;M}HyU:-wȋy \0% V ;Y"x^cm^Ŭ\<9(\nHA dO\>яnD-pmw V@(QC("SSp(pRD`H)`at9t"vF1pӶ5+.krjh$*Χ\\;I+f0A)ixq\|#"NB\nR) @1\|\>&kOZ14p\0\0 IQgY)&?bH\\k! g/6rIe\lj36Mqzsf8͙rr ;\tNnZ=O}'yr5AP. uC!ЀE1Q4(@\0;>|gif>|*5/8|*5/8||||> We saved it as comma separated values. <\input| >> a=read.table("example2.csv",head=T,row.names=1,as.is=3, sep=",") <\input| >> a <\output> \ \ \ \ \ \ \ \ height \ \ \ \ \ \ \ \ \ \ \ \ email hair.color age michael \ \ \ 175 \ \ \ \ \ \ \ \ \ lachmann \ \ \ \ \ brown \ NA sarah \ \ \ \ \ 180 \ \ \ \ sarah@aol.com \ \ \ \ \ black \ 20 fred \ \ \ \ \ \ 150 \ \ f.smith@msn.net \ \ \ \ \ \ \ \ \ \ \ \ 10 John \ \ \ \ \ \ 210 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ blond \ 25 Bilbo \ \ \ \ \ \ 90 b.bagins@shire.me \ \ \ \ \ black 111 Andrew \ \ \ \ 150 \ \ \ \ \ \ andy@aol.de \ \ \ \ \ blond \ 35 <\input| >> \; > Normaly, R will convert a string column that is read in into categorical data. To prevent R from doing that, you should give the argument, with the columns that should be read as strings. If a column has values that are essencially all different, store as string. If some values are the same and you will compare between values, use categorical data (i.e. no ) The are many ways to access the data in the table: <\input| >> a$height <\output> [1] 175 180 150 210 \ 90 150 <\input| >> a$hei <\output> [1] 175 180 150 210 \ 90 150 <\input| >> a$hai <\output> [1] brown black \ \ \ \ \ \ blond black blond Levels: \ black blond brown <\input| >> \; > The method using the dollar sign takes a column. You can see that you can abbreviate the name. <\input| >> a[,1] <\output> [1] 175 180 150 210 \ 90 150 <\input| >> a[1:2,] <\output> \ \ \ \ \ \ \ \ height \ \ \ \ \ \ \ \ email hair.color age michael \ \ \ 175 \ \ \ \ \ lachmann \ \ \ \ \ brown \ NA sarah \ \ \ \ \ 180 sarah@aol.com \ \ \ \ \ black \ 20 <\input| >> a[,"email"] <\output> [1] "lachmann" \ \ \ \ \ \ \ \ \ "sarah@aol.com" \ \ \ \ "f.smith@msn.net" \ [4] "" \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ "b.bagins@shire.me" "andy@aol.de" \ \ \ \ \ <\input| >> a[,"ema"] <\output> Error in "[.data.frame"(a, , "ema") : undefined columns selected <\input| >> \; > You can not abreviate when you use . We want to sort the table by height. How do we do that? Here the function helps. It gives us the order of elements. <\input| >> a <\output> \ \ \ \ \ \ \ \ height \ \ \ \ \ \ \ \ \ \ \ \ email hair.color age michael \ \ \ 175 \ \ \ \ \ \ \ \ \ lachmann \ \ \ \ \ brown \ NA sarah \ \ \ \ \ 180 \ \ \ \ sarah@aol.com \ \ \ \ \ black \ 20 fred \ \ \ \ \ \ 150 \ \ f.smith@msn.net \ \ \ \ \ \ \ \ \ \ \ \ 10 John \ \ \ \ \ \ 210 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ blond \ 25 Bilbo \ \ \ \ \ \ 90 b.bagins@shire.me \ \ \ \ \ black 111 Andrew \ \ \ \ 150 \ \ \ \ \ \ andy@aol.de \ \ \ \ \ blond \ 35 <\input| >> order(a$height) <\output> [1] 5 3 6 1 2 4 <\input| >> \; > So will be sorted by height if we write first the 5th row, then the 3rd, then the 6th, then the 1st, and so on. we can do that like this: <\input| >> i=order(a$height) <\input| >> i <\output> [1] 5 3 6 1 2 4 <\input| >> a[i,] <\output> \ \ \ \ \ \ \ \ height \ \ \ \ \ \ \ \ \ \ \ \ email hair.color age Bilbo \ \ \ \ \ \ 90 b.bagins@shire.me \ \ \ \ \ black 111 fred \ \ \ \ \ \ 150 \ \ f.smith@msn.net \ \ \ \ \ \ \ \ \ \ \ \ 10 Andrew \ \ \ \ 150 \ \ \ \ \ \ andy@aol.de \ \ \ \ \ blond \ 35 michael \ \ \ 175 \ \ \ \ \ \ \ \ \ lachmann \ \ \ \ \ brown \ NA sarah \ \ \ \ \ 180 \ \ \ \ sarah@aol.com \ \ \ \ \ black \ 20 John \ \ \ \ \ \ 210 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ blond \ 25 <\input| >> a[ order( a$height ) , ] <\output> \ \ \ \ \ \ \ \ height \ \ \ \ \ \ \ \ \ \ \ \ email hair.color age Bilbo \ \ \ \ \ \ 90 b.bagins@shire.me \ \ \ \ \ black 111 fred \ \ \ \ \ \ 150 \ \ f.smith@msn.net \ \ \ \ \ \ \ \ \ \ \ \ 10 Andrew \ \ \ \ 150 \ \ \ \ \ \ andy@aol.de \ \ \ \ \ blond \ 35 michael \ \ \ 175 \ \ \ \ \ \ \ \ \ lachmann \ \ \ \ \ brown \ NA sarah \ \ \ \ \ 180 \ \ \ \ sarah@aol.com \ \ \ \ \ black \ 20 John \ \ \ \ \ \ 210 \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ blond \ 25 <\input| >> \; > In R, almost everything can have names: <\input| >> a=1:4 <\input| >> a <\output> [1] 1 2 3 4 <\input| >> names(a) <\output> NULL <\input| >> names(a)=c("a","b","c","d") <\input| >> a <\output> a b c d\ 1 2 3 4\ <\input| >> names(a) <\output> [1] "a" "b" "c" "d" <\input| >> \; > Now we see that each entry has a name. An easier way to do the same: <\input| >> x=c(first=3,second=4) <\input| >> x <\output> \ first second\ \ \ \ \ \ 3 \ \ \ \ \ 4\ <\input| >> \; > This is useful when functions return values, because it tells us what each value is: <\input| >> a=rnorm(100) <\input| >> summary(a) <\output> \ \ \ Min. 1st Qu. \ Median \ \ \ Mean 3rd Qu. \ \ \ Max.\ -2.2070 -0.6992 \ 0.2073 \ 0.1538 \ 0.9701 \ 2.6370\ <\input| >> \; > is a function, and the values that it returns have names: <\input| >> sum.a=summary(a) <\input| >> names(sum.a) <\output> [1] "Min." \ \ \ "1st Qu." "Median" \ "Mean" \ \ \ "3rd Qu." "Max." \ \ <\input| >> sum.a["Mean"] <\output> \ \ Mean\ 0.1538\ <\input| >> sum.a["Max."] <\output> \ Max.\ 2.637\ <\input| >> \; > matrices and data.frames have names for the rows and columns: <\input| >> a=matrix(1:6,2,3) <\input| >> a <\output> \ \ \ \ \ [,1] [,2] [,3] [1,] \ \ \ 1 \ \ \ 3 \ \ \ 5 [2,] \ \ \ 2 \ \ \ 4 \ \ \ 6 <\input| >> rownames(a)=c("first row","second row") <\input| >> colnames(a)=c("a","b","c") <\input| >> a <\output> \ \ \ \ \ \ \ \ \ \ \ a b c first row \ 1 3 5 second row 2 4 6 <\input| >> a[,"a"] <\output> \ first row second row\ \ \ \ \ \ \ \ \ \ 1 \ \ \ \ \ \ \ \ \ 2\ <\input| >> a[,"b"] <\output> \ first row second row\ \ \ \ \ \ \ \ \ \ 3 \ \ \ \ \ \ \ \ \ 4\ <\input| >> \; > Row names are especially important when we want to use data from different sources: for example when we have results of one experiment for some genes, and more results of from a second experiment. The row names allow us then to quickly connect the results for the same genes. A matrix basically is a vector, and as such can hold only one type of data. This can cause strange results sometime: <\input| >> x=matrix(1:6,2,3) <\input| >> x <\output> \ \ \ \ \ [,1] [,2] [,3] [1,] \ \ \ 1 \ \ \ 3 \ \ \ 5 [2,] \ \ \ 2 \ \ \ 4 \ \ \ 6 <\input| >> x[2,2]="four" <\input| >> x <\output> \ \ \ \ \ [,1] [,2] \ \ [,3] [1,] "1" \ "3" \ \ \ "5"\ [2,] "2" \ "four" "6"\ \; > by changing a single entry in x, the whole matrix was changed to strings. <\input| >> x=data.frame(a=1:3,b=4:6) <\input| >> x <\output> \ \ a b 1 1 4 2 2 5 3 3 6 <\input| >> x[2,2] <\output> [1] 5 <\input| >> x[2,2]="five" <\input| >> x <\output> \ \ a \ \ \ b 1 1 \ \ \ 4 2 2 five 3 3 \ \ \ 6 <\input| >> x[,2] <\output> [1] "4" \ \ \ "five" "6" \ \ <\input| >> x[,1] <\output> [1] 1 2 3 <\input| >> \; > You can see that only one column of x was changed to strings. A data.frame is a list of vectors (or matrices), all of which have the same length, and can thus be addressed as an array. <\input| >> x[1] <\output> \ \ a 1 1 2 2 3 3 <\input| >> x[2] <\output> \ \ \ \ \ b 1 \ \ \ 4 2 five 3 \ \ \ 6 <\input| >> \; > Last time we talked a bit about functions. Now some more. Let us define a simple function: <\input| >> f=function(x) x+1 <\input| >> f(4) <\output> [1] 5 <\input| >> f(9.3) <\output> [1] 10.3 <\input| >> x=17 <\input| >> f(2) <\output> [1] 3 <\input| >> x <\output> [1] 17 <\input| >> \; > So, a function is very easy to define: We just say followed by an expression. <\input| >> square= function(z) z^2 <\input| >> square(3) <\output> [1] 9 <\input| >> square(1:5) <\output> [1] \ 1 \ 4 \ 9 16 25 <\input| >> square <\output> function(z) z^2 > We can see that functions can simply be stored in variables, like numbers or strings. Functions can also take several arguments: <\input| >> mult = function(x,y) x*y <\input| >> mult(2,3) <\output> [1] 6 <\input| >> mult(1:4,2:5) <\output> [1] \ 2 \ 6 12 20 <\input| >> \; > Applying a function to a vector <\input| >> a=1:10 <\input| >> f=function(x) x+1 <\input| >> sapply(a,f) <\output> \ [1] \ 2 \ 3 \ 4 \ 5 \ 6 \ 7 \ 8 \ 9 10 11 <\input| >> \; > applies the function to each element in the vector, and returns the result. If the result is a vector, then we get a matrix: <\input| >> g=function(x) c(x,x+1) <\input| >> g(2) <\output> [1] 2 3 <\input| >> g(5) <\output> [1] 5 6 <\input| >> sapply(1:4,g) <\output> \ \ \ \ \ [,1] [,2] [,3] [,4] [1,] \ \ \ 1 \ \ \ 2 \ \ \ 3 \ \ \ 4 [2,] \ \ \ 2 \ \ \ 3 \ \ \ 4 \ \ \ 5 <\input| >> \; > Sometimes we need more than one expression for the calculation. In that case we can enclose the calculations with . The result will be the last expression. <\input| >> die.roll.6=function(){ x=sample(1:6,1); x==6 } <\input| >> die.roll.6() <\output> [1] FALSE <\input| >> \; > Assignments in functions do not affect the outside: <\input| >> x=100 <\input| >> die.roll.6() <\output> [1] FALSE <\input| >> x <\output> [1] 100 <\input| >> \; > To see what is happenening in a function, we can add print statements: <\input| >> die.roll.6=function(){ x=sample(1:6,1); print(x); x==6 } <\input| >> die.roll.6() <\output> [1] 2 [1] FALSE <\input| >> \; <\input| >> \; > R also has conditional statements, and loops: <\input| >> f = function(x) { if( x \ 3) 4 else 5} <\input| >> f(2) <\output> [1] 5 <\input| >> f(6) <\output> [1] 4 <\input| >> \; > <\input| >> for(i in 1:10) print(i) <\output> [1] 1 [1] 2 [1] 3 [1] 4 [1] 5 [1] 6 [1] 7 [1] 8 [1] 9 [1] 10 <\input| >> a=1 <\input| >> for(i in 1:10) a = a * i <\input| >> a <\output> [1] 3628800 \; > has the following structure: you give a variable that will iterate over a vector or list of things.\ The expression will be called with each i. \; \; <\initial> <\collection> <\references> <\collection> |?>> |?>> |?>> |?>> |?>> |?>> |?>> |?>> <\auxiliary> <\collection> <\associate|toc> Lecture 3 |read.table> |names> |What is the difference between matrices and data.frames?> |Functions> |Conditionals> |Loops>