How to find unique factor names by groups

Question

I'm looking to find unique values : species (here it is species code called "SPID", with the first four letters from genus name first four letters from species name) - in different groups of my data frame : habitats (there is 3 habitats (variable "hab"), named TA, TB and TC).

Here is an output of my data :

library(dplyr)

brk%>%
+   dplyr::select(spid,hab)%>%
+   dplyr::sample_n(20)%>%
+   dput()
structure(list(spid = structure(c(127L, 78L, 33L, 162L, 81L, 
72L, 72L, 196L, 196L, 241L, 240L, 238L, 190L, 181L, 188L, 162L, 
214L, 13L, 178L, 78L), .Label = c("ACROEMER", "ACROMEGA", "AEROSUBPM", 
"AMAZDIPL", "ANASAURI", "ANASPILI", "ANDRABER", "ANDRBILO", "ANEULATI", 
"BAZZDECR", "BAZZDECRM", "BAZZMASC", "BAZZNITI", "BAZZPRAE", 
"BAZZROCA", "BRACEURY", "BUCKMEMB", "CALYARGU", "CALYFISS", "CALYMASC", 
"CALYPALI", "CALYPERU", "CAMPARCTM", "CAMPAURE", "CAMPCRAT", 
"CAMPFLEX", "CAMPJAME", "CAMPROBI", "CAMPTHWA", "CEPHVAGI", "CERABELA", 
"CERACORN", "CERAZENK", "CHEICAME", "CHEICORDI", "CHEIDECU", 
"CHEIMONT", "CHEISERP", "CHEISURR", "CHEITRIF", "CHEIUSAM", "CHEIXANT", 
"COLOCEAT", "COLOHASK", "COLOHILD", "COLOOBLI", "COLOPEPO", "COLOTANZ", 
"COLOZENK", "COLUBENO", "COLUCALY", "COLUDIGI", "COLUHUMB", "COLUOBES", 
"COLUTENU", "CONOTRAP", "CRYPMART", "CUSPCONT", "CYCLBORB", "CYCLBREV", 
"CYLIKIAE", "DALTANGU", "DALTLATI", "DENDBORB", "DICRBILLB", 
"DIPLCAVI", "DIPLCOGO", "DIPLCORN", "DREPCULT", "DREPHELE", "DREPMADA", 
"DREPPHYS", "ECTRREGU", "ECTRVALE", "FISSASPL", "FISSMEGAH", 
"FISSSCIO", "FRULAPIC", "FRULAPICU", "FRULBORB", "FRULCAPE", 
"FRULGROS", "FRULHUMB", "FRULLIND", "FRULREPA", "FRULSCHI", "FRULSERR", 
"FRULUSAMR", "FRULVARI", "FUSCCONN", "GOTTNEES", "GOTTSCHI", 
"GOTTSPHA", "GROULAXO", "HAPLSTIC", "HERBDICR", "HERBJUNI", "HERBMAUR", 
"HETEDUBI", "HETESPLE", "HETESPN", "HOLOBORB", "HOLOCYLI", "HYPNCUPR", 
"ISOPCHRY", "ISOPCITR", "ISOPINTO", "ISOTAUBE", "JAEGSOLI", "JAEGSOLIR", 
"KURZCAPI", "KURZCAPIS", "LEJEALAT", "LEJEANIS", "LEJECONF", 
"LEJEECKL", "LEJEFLAV", "LEJELOMA", "LEJEOBTU", "LEJERAMO", "LEJETABU", 
"LEJETUBE", "LEJEVILL", "LEPIAFRI", "LEPICESP", "LEPIDELE", "LEPIHIRS", 
"LEPISTUH", "LEPISTUHP", "LEPTFLEX", "LEPTINFU", "LEPTMACU", 
"LEUCANGU", "LEUCBIFI", "LEUCBORY", "LEUCCANDI", "LEUCCAPI", 
"LEUCCINC", "LEUCDELI", "LEUCGRAN", "LEUCHILD", "LEUCISLE", "LEUCLEPE", 
"LEUCMAYO", "LEUCSEYC", "LOPHBORB", "LOPHCOAD", "LOPHCONC", "LOPHDIFF", 
"LOPHEULO", "LOPHMULT", "LOPHMURI", "LOPHNIGR", "LOPHSUBF", "MACRACID", 
"MACRMAUR", "MACRMICR", "MACRPALL", "MACRSERP", "MACRSULC", "MACRTENU", 
"MASTDICL", "METZCONS", "METZFURC", "METZLEPT", "METZMADA", "MICRAFRI", 
"MICRANKA", "MICRDISP", "MICRINFL", "MICRKAME", "MICROBLO", "MICRSTRA", 
"MITTLIMO", "MNIOFUSC", "PAPICOMP", "PLAGANGU", "PLAGDREP", "PLAGPECT", 
"PLAGRENA", "PLAGREPA", "PLAGRODR", "PLAGTERE", "PLEUGIGA", "PLICHIRT", 
"POLYCOMM", "POROELON", "POROMADA", "POROUSAG", "PRIOGRAT", "PSEUDECI", 
"PTYCSTRI", "PYRRSPIN", "RACOAFRI", "RADUANKE", "RADUAPPR", "RADUBORB", 
"RADUBORY", "RADUCOMO", "RADUEVEL", "RADUFULV", "RADUMADA", "RADUSTEN", 
"RADUTABU", "RADUVOLU", "RHAPCRIS", "RHAPGRAC", "RHAPRUBR", "RICCAMAZ", 
"RICCEROS", "RICCFAST", "RICCLIMB", "RICCLONG", "SCHLBADI", "SCHLMICRO", 
"SCHLOANGU", "SCHLSQUA", "SEMACRAS", "SEMASCHI", "SEMASUBP", 
"SERPCYRT", "SOLEBORG", "SOLEONRA", "SOLESPHA", "SPHATUMI", "SPHEMINU", 
"SYRRAFRI", "SYRRAPER", "SYRRDIMO", "SYRRGAUD", "SYRRHISP", "SYRRPOTT", 
"SYRRPROL", "SYRRPROLA", "SYZYPURP", "TAXICONFO", "TELACOAC", 
"TELADIAC", "TELANEMA", "TRICADHA", "TRICDEBE", "TRICPERV", "ULOTFULV", 
"WARBLEPT", "ZYGOINTE", "ZYGOREIN"), class = "factor"), hab = structure(c(1L, 
3L, 3L, 2L, 3L, 2L, 1L, 2L, 3L, 1L, 2L, 2L, 3L, 1L, 2L, 1L, 3L, 
2L, 3L, 2L), .Label = c("TA", "TB", "TC"), class = "factor")), row.names = c(NA, 
-20L), class = "data.frame")

I tried with :

dplyr::select(spid,hab)%>%
 dplyr::group_by(hab)%>%
 dplyr::summarise(n_distinct(spid))

Obviously it won't show me what I want. How could I find the name of unique species by habitat ?

Thank you for your help,

Answer 1

EDITED to match ... A very low key solution but sometimes tables are nice ways of displaying this...

table(your_data$spid, 
      your_data$hab)[rowSums(table(your_data$spid, 
                                   your_data$hab)) == 1, ]


#>           
#>            TA TB TC
#>   BAZZNITI  0  1  0
#>   CERAZENK  0  0  1
#>   FRULCAPE  0  0  1
#>   LEPIHIRS  1  0  0
#>   PLAGDREP  0  0  1
#>   PLAGREPA  1  0  0
#>   POROMADA  0  1  0
#>   PRIOGRAT  0  0  1
#>   SCHLBADI  0  0  1
#>   TELADIAC  0  1  0
#>   TRICADHA  0  1  0
#>   TRICDEBE  1  0  0

Answer 2

You could try:

df <- dplyr::select(brk, spid, hab)

lapply(split(df, df$hab), unique)

Which will give you a list containing one data frame for each habitat with the unique species in the spid column:

$TA
       spid hab
1  LEPIHIRS  TA
7  DREPPHYS  TA
10 TRICDEBE  TA
14 PLAGREPA  TA
16 MASTDICL  TA

$TB
       spid hab
4  MASTDICL  TB
6  DREPPHYS  TB
8  RADUAPPR  TB
11 TRICADHA  TB
12 TELADIAC  TB
15 POROMADA  TB
18 BAZZNITI  TB
20 FRULAPIC  TB

$TC
       spid hab
2  FRULAPIC  TC
3  CERAZENK  TC
5  FRULCAPE  TC
9  RADUAPPR  TC
13 PRIOGRAT  TC
17 SCHLBADI  TC
19 PLAGDREP  TC

---

Edit based on clarification:

df2 <- as.data.frame(table(df$spid, df$hab)[rowSums(table(df$spid, df$hab))==1,])
df2[df2$Freq != 0,]
#>        Var1 Var2 Freq
#> 4  LEPIHIRS   TA    1
#> 6  PLAGREPA   TA    1
#> 12 TRICDEBE   TA    1
#> 13 BAZZNITI   TB    1
#> 19 POROMADA   TB    1
#> 22 TELADIAC   TB    1
#> 23 TRICADHA   TB    1
#> 26 CERAZENK   TC    1
#> 27 FRULCAPE   TC    1
#> 29 PLAGDREP   TC    1
#> 32 PRIOGRAT   TC    1
#> 33 SCHLBADI   TC    1

Answer 3

Here's a revised solution to show spid's associated with 1 hab only.

> brk %>% 
    group_by(spid) %>% 
    summarize(nn = n_distinct(hab)) %>% 
    filter(nn == 1) %>%
    ungroup()

# A tibble: 12 x 2
   spid        nn
   <fct>    <int>
 1 BAZZNITI     1
 2 CERAZENK     1
 3 FRULCAPE     1
 4 LEPIHIRS     1
 5 PLAGDREP     1
 6 PLAGREPA     1
 7 POROMADA     1
 8 PRIOGRAT     1
 9 SCHLBADI     1
10 TELADIAC     1
11 TRICADHA     1
12 TRICDEBE     1

Answer 4

dplyr solution:

brk %>% 
  distinct(spid, hab) %>% 
  arrange(hab)

              spid hab
1  LEPIHIRS  TA
2  DREPPHYS  TA
3  TRICDEBE  TA
4  PLAGREPA  TA
5  MASTDICL  TA
6  MASTDICL  TB
7  DREPPHYS  TB
8  RADUAPPR  TB
9  TRICADHA  TB
10 TELADIAC  TB
11 POROMADA  TB
12 BAZZNITI  TB
13 FRULAPIC  TB
14 FRULAPIC  TC
15 CERAZENK  TC
16 FRULCAPE  TC
17 RADUAPPR  TC
18 PRIOGRAT  TC
19 SCHLBADI  TC
20 PLAGDREP  TC

Edit Based on clarification

brk %>% 
  group_by(spid) %>% 
  mutate(n = n()) %>% 
  filter(n == 1) %>% 
  select(-n)

# A tibble: 12 x 2
# Groups:   spid [12]
   spid     hab  
   <fct>    <fct>
 1 LEPIHIRS TA   
 2 CERAZENK TC   
 3 FRULCAPE TC   
 4 TRICDEBE TA   
 5 TRICADHA TB   
 6 TELADIAC TB   
 7 PRIOGRAT TC   
 8 PLAGREPA TA   
 9 POROMADA TB   
10 SCHLBADI TC   
11 BAZZNITI TB   
12 PLAGDREP TC

Answer 5

Thank you for yours answers. Yes it's kind of what I want, but do you know why a species (spid) can be unique in 2 habitats ? What I want is to show all species that are only present in 1 habitat, maybe I didn't explained well, sorry if it's the case.